From an accidental search regarding fiber you are overall becoming an enjoyable TH-camr, with passion, entertainment, and your entrepreneurial spirit is inspiring me to do my own thing. I’m so excited for your router. I’ll probably be one of the first. I’m ready to replace my ubiquiti gear.
@@tomazzaman Keep it up! It’s impressive watching your growth, even the last few weeks. Your videos feel like you’ve been producing for years. You also encapsulate entire topics to clear undertake parts, and English isn’t even your first language! Have a great 2024! Excited to see you take off! We’ll all be here for the ride!
I currently own a service at my company and have dealt with this same issue. Typically more often than not, the customers who think the product is too expensive don't understand the value of what they are receiving. In my case, I usually ensure they understand the service and compare it to competitive services as well as what it would cost them to hire someone to provide themselves with the service. Thanks for the great video! Keep up the good work!
From hacking his ISP's router to this. Love the content. My original answers were made before watching the BOM video, and seeing it I kinda wanted to revise my answers seeing how much it was likely to cost. I believe $300 plus whatever the margin would be seems fair honestly.
One thing at is (yet) missing from that market research is to determine what would be the target comparable existing products and what is their average sale price. Comparison should look at features, performance and warranty. For example, if all comparable products cost a little more than 500$, then people looking for that product will still be attracted to it at 500$ because it is still cheaper than all the alternatives in order to get the same value. In order to determine the comparable products, a performance benchmark of the total throughput should help. In my opinion, people who only expect to pay 250$ for a router want extra features that they probably would not use, but still say "it's a nice to have" for sometime in the future without having an actual plan to start using them. People with an actual "business" need for the extra feature are the one that know how much that feature is worth. I wonder how many people from that survey do not even have a 10G capable network and have no plan on upgrading... On a purely business point of view, you would probably make as much profit by serving the top 20% higher responders than the mass mid-range.
It's less about the features and more about how much pain and what types of it the different market segments are willing to accept. One segment may be willing to spend thousands of dollars on a 10gig router while another is only willing to do a few hundred. Why such a difference in price for the same feature? it comes down to what pain points they experience the most. Examples: *Government:* Major pain points are organizational. It's a PITA to get any new vendor authorized to work within a governmental agency. Willing to blow a lot of money to not have to deal with their own bureaucracy. Normally siloed into a few vendors. *Large Enterprise:* Similar to government, but may focus more on "scale-out" and remote management/automation features. Similar pain points but with a bit more wiggle room. Still won't deal with small vendors very often. *Small-Medium Enterprise:* Pain points start to focus on cost/benefit relationship of products used. Similar requirements to large enterprise needs for management and uniformity, but more often willing to take risks on unknown products in order to get a better TCO. *Small business:* Pain point is money and uptime. Quite often they'll be willing to cut back on features in order to find something stable and consistent. Quite often they also don't have many manhours available to focus on technology and instead need higher usability requirements compared to SME. *Owner/Operator Business:* Money and uptime, but to lesser degrees. Generally they want to use off-the-shelf parts that they have to think about once and never again. Since they don't have many/any employees but themselves they'll be willing to put up with downtime as long as the cost is low enough. Also quite often they'll hire someone else to set up their equipment and actually don't make the purchasing decisions themselves. The people they hire likely have pain points similar to SME or Geeks, depending on who they hire. *Geeks:* People who love tech, quite often spend large amount of money on products as status symbols or because it's the new thing. Early adopters, their pain points are individual but ease of use and product appearance needs are high for much of the market segment. *Homelabbers:* Geeks but they operate more like an SME. They want high-end equipment for cheap running large feature subsets since they want to learn about technology in-depth. Quite often willing to put up with large amounts of technical and stability pain as long as the product is performant. Primary pain point is usually cost (even though they may spend thousands more than a "Geek" would). As it stands, the segments this new router is best oriented towards are the small business and geek segments. Maybe SME and homelabbers, however: -As a homelabber I'd be aiming at something like the CCR2004-1G-12S+2XS from Mikrotik at the $500 price point, or purchase used enterprise equipment. -As an SME I'd be aiming at something from second-tier companies (Ubiquiti, Ruckus, Aruba, Dell), Juniper or Cisco. All of those products have some form of remote or cloud management service and offer a lot of capability and stability at a decent price. They also provide uniformity for the network with brands likely to still be selling products in 10-20 years' time. Just found this product so I'll need to look into this router more indepth before I say whether I'd buy it or not. But, that's my general thoughts.
@@jasonmonsen7052 I love you answer. This new router being essentially a piece of hardware, it does not do anything by itself. It needs a firewall software running on it to create its value. If the targeted softwares are OPNSense, PFSense and VyOS, then the implicit market segment for the hardware will be the same ones as the software that will be running on it, or maybe a subset of them. This is probably what was missing in the survey that gave such a wide range of expected sale price between the responders: knowing what market each responder were identifying themselves to. The geeks or owner/operator were surely not answering the same way as someone responding with a large business in mind.
@@viaujoc Your first couple sentences leads into what I was thinking about yesterday after I posted. Tomaz' follow up video about adding value will likely cover this, but one of the things that could set this product apart is a dead-simple out-of-the-box experience for the router itself. I don't know how hard it'd be to create some wizards within the mentioned OSes, but if you could offer the raw power while also offering an experience similar to an off-the-shelf Linksys box, I think it'd be a winner. Some unique value-adding features to the software could be a strong selling point. Thinking about what group an easy to use powerhouse would appeal to, I think of: gamers, streamers, and any small company working with CAD, video, or AI systems. Dual 10gip with 4 2.5gbps ports would be pretty nice for anyone working with large data in a hybrid setup (cloud and on-premises), or small offices with off-premises cloud infrastructure. Tomaz being a YTer likely has some good ideas around what advanced features would be useful for streamers. Things like QoS or VLANs could be very useful to creators who could benefit from it but might not have enough experience to know what they're doing. Including some newbie-friendly guides that go over some creator-centric features which might be useful to them (and giving examples of when or why) and step-by-step instructions (or wizards) to add these features would help a lot. For more technical "geeks/homelabbers" that value-add might not be as useful, but it wouldn't detract from the product itself as long as there's an advanced mode. I did mention a good Mikrotik alternative at near the same price, however it is a weaker processor and only has 12x1gip ports alongside its 2x10gip, so this one could make sense in certain network configurations. For gamers, while this switch would be overkill, 2.5gip NICs are starting to show up on motherboards and I could see myself justifying the purchase. As another commenter mentioned on this video, most "gaming" routers are complete crap and a money grab, offering a legitimate "prosumer" option would be something fairly unique. I wouldn't focus on selling to businesses at first with the version 1 (outside streamers and some large dataset studios). If this does see multiple revisions and becomes something longer-term then I could see businesses considering it. The port layout is pretty great for a flashable router and would beat out a lot of the x86 whitebox competition.
@@viaujoc Just a heads up your last post is hidden for some reason. And my last answer was kind of talking "at" you instead of "to" you. You're right though that it makes a lot of sense to understand how people were approaching their dollar values. I don't think many would be coming at this from a large business point of view (I already have vendor contracts), but it is still very interesting to know what they think their use case for such a router would be.
@@jasonmonsen7052 Unfortunately integrating third party open source softwares is not an easy feat. Some softwares (ex: pfSense CE) have explicit restrictions in their licence agreement prohibiting selling a hardware with their software pre-installed. However, additional value may be created by giving the buyer some guaranty that a specific third party software product runs stably on it by thoroughly testing it and providing installation guidelines and best practices to integrate the hardware and software together. PCEngines did that and had a respectable product with a good sales volume for many years. I also agree with you that targeting medium to large businesses on day 1 may not be a good idea because those companies will usually be looking for maintenance contracts and strict response time commitments. This involves having a dedicated support team with the salaries and other related expenses. Nice discussion we have here. I hope this will be useful to Tomaz...
One other thing I'd like to mention, power efficiency. Most of the time hardware is idling, so if the idle power consumption can be optimised it would help with those of us paying crazy amounts for power.
Aye. The CPU we chose is a very efficient one, single digit watts while idle. Once we get the development kit, I’ll make a dedicated video just about this topic.
I would try to make the router look more expensive w/o actually increasing it's cost (much). I also wouldn't just show them a router, I would explain its benefits over the competition.
Agreed. I do have a slight preference for aluminum over plastic, but formed sheet is fine. CNC hogouts seems like gross overkill. And, I have plenty of plastic stuff too, so even that's not an insurmountable barrier. I do industrial controls, so while I'd like one for my own use, to build one into equipment I sell, I'd like DIN rail mount in bookshelf orientation, actual terminals for power vs. a barrel connector (preferably 24VDC) , and all the lights and ports on the exposed edge.
Excellent! Perceived value was one of the tools British Airways used to price up the cost of tickets for Concorde flights to the US - finally turning a profit after years of losses by *increasing* the ticket prices!
This video made me go back and check your other videos about your router plans. I'll be watching this with interest. I recently watched a video on "Serve the Home" channel, the video is titled "The EVERYTHING $300 Fanless Home Server". Since you're figures were around the same I would suggest looking at that hardware and maybe thinking about what you can copy in terms of BOM and also differentiate in terms of features.
If you're thinking about shipping globally, have you considered asking people about regional price sensitivity? Different countries have differently-scaled economies after all, and that will affect how much people can afford to pay for anything.
Awesome video Tomaž! It is very helpful for the ones of us, who have little idea what goes into the pricing parts of product development! 😄 Ži čakam na nasledni videjo
loving the idea and the drive to get a good open-source router on to the market! I put my own router together using a Lenovo mini PC, as I couldn't find what I wanted on the market. One nice to have feature, if I was looking at building a router from scratch, a dedicated management port that can accept POE+ to power the unit. This allows the provision of redundant power if needed, or just makes an install much cleaner :) That said, your design ticks all the wish list items on what I would want on a router, multiple SFP+, 2.5GB LAN, good IDS/IPS performance, a wide choice of OS. Keep up the awesome work!
i would definitely love to see a video comparing your product to the competition. maybe an average isp modem vs an average mid range router vs an average high end one on the market right now.
My solution would be to offer the router in variations that cut down on some of the features less useful for an average user - like having 64 GB of onboard storage... In a router?.. And lower the price point for that variation accordingly. That way, you could get both the customers who believe the current iteration of your router is too expensive (likely because they won't be making use of a lot of its features) and the ones who are willing to pay what it's actually worth (likely the people who appreciate all of its abilities in full).
For the challenge- I would ask myself whether the respondents to the survey really are the ideal customer I’d want to sell to. If not, I’d run the survey again and try to get a lot more targeted about who I try to get the survey in front of. If so, I’d do a few different things: see if I could reduce COGS, and since that’s unlikely without significantly sacrificing quality, make sure I thoroughly educate the potential customer before presenting the price. Another idea could be to offer a lower quality product (1G networking or maybe 2.5?) for the people who want a lower price point. Could even skip the aluminum case on that one, or use an aluminum case that’s industrially available for mini-pcs and possibly cut your costs that way. Considering the fact that some of the Protectli Vaults go for $500-$700, I think this is excellent value for what it is, especially with 10G.
Hi Tomaz...a great video! You have gotten us to think about how things are made and priced. Looking forward to seeing where this goes. I think you are on the right track!
Make a tiered router= make a pcb that if fully populated has every option ( al the GB/2.5G & 2x SFP+ caged) and a nice anodized alu case (400-500€ range). When only populated with 1x sfp+ & 1x 2.5G & 1x GB port and a sheet steel housing then it's the 200€ range (minimum edition). You can do a H4X0r edition to: here's the pcb and components have fun ;) or a Lite editon = same as minimum but no case.
I'm surprised how many have unrealistic expectations to what a product like this cost. I'm guessing they are either not very well versed in the networking world or are from countries with a much lower per capita income than the average. I consider this product to be a prosumer product or SOHO if you will. And based on the current specs, compared to similar products, I think the "founders edition" fits nicely somewhere in the 500-600 USD range, while the cheaper one (plastic instead of aluminium) in the 500-550 USD range. I have a router from Netgate, so I'm not looking to replace it yet, but it's interesting to follow this journey. It's both entertaining and educational. The quality, both in terms of content and presentation, of this channel is awesome, keep up the impressive work Tomaž :D
Its most likely regional based. I mentioned this in a comment in previous video. For example in Poland which is a poor country, the most similar to this router devices are: Ubiquiti X-SFP for 429 PLN (106 USD) QNAP QSW-2104-2S for 629 PLN (156 USD)
I initially searched different stuff and watched your other videos, but you got my interest with announcements of building your own router I wish you good luck, you got me interested in end product. One thing I seams to miss is what is your intention for target mass of buyers, business/home or ??? Especially as I am trying to figure out what is your answer on the same 4 questions you wanted from your subscribers in order to get price point. Hope you will clear few of these questions in next videos, because target usage would be one of the turning point for some target buyers, me included. eg. ground up design of your LAN/WLAN or you are searching for expansion or step to evolve your existing infrastructure which greatly influence how much you pay for a device and if it will fill your purpose. Would be glad to talk about it with you and your subscribers. Hope will see replies.
Thanks for the kind words! I'm currently focusing on the advanced home users because I'm one of them, but do have plans to address a couple of other markets in the future! I do plan to make a video about that as well, probably not just yet, but when things really clear out.
@@tomazzaman That is the answer I wanted to hear, because like you I am advanced home user, very interested in your product which from my brief analyze fits exactly my needs for mid range network infrastructure complexity and future proof that would allow me to archive my home plan and level of independency and control over it. Hvala, veoma uživam u tvojim videima.
It also sounds like you can use the price sensitivity data to produce a graph of "expected profit" by taking in the raw cost, subtracting the price point, and multiplying by expected percentage of clients. Then, use that to maximize profit (without going to the extremes, of course). I wonder if that's a reasonable approach.
It is, but the PSM gives you more of a general idea than a final one. For example, imagine you produce the unit overseas. This means you also have to ship it to the warehouse from which you sell. This also adds costs, that are not in BOM. And there’s plenty of costs like this that are hard to account for this far in advance.
After a long doom scroll i came across you, been hooked since. Will buy the second it hits the market, i need to get rid of the prop HW our ISPs in India give.
Have you considered crowd sourcing some models that can be 3D printed at home for the case and selling the the rest in a kit? Not to mention people buying “high performance routers” likely want a way to easily mount it in their network rack. Any thoughts?
A highly intriguing project & amazingly entertaining videos. And for me, watching every next video is also adding to an increase of the perceived value and/or willingness to maybe pay more for the unit. (I previously filled in the Pricing form)
Good luck. Lot of money for normal people and us nerds know hoe to buy better kit for less$ small network appliances is a pretty saturated marketplace, good luck beating some of cwwk's latest.
Great video. In the next video could you discuss what operating systems we'll be able to run on this ARM in more specific detail? For example, will there actually be an ARM build of OPNsense built and made available? - I know pfSense has an ARM build but it's currently restricted to their own hardware appliances and I don't believe OPNsense has an ARM build yet.
I have it planned, but not in the next one, I have a couple of other videos to make before we get to the actual software and hardware, primarily because I don't have the development kit yet and also because I want to share the full journey and not skip any steps. Thanks!
id like to think i had a hand in changeing out the one 10GbE to another SFP+ Port as i suggested, and i will continue to watch as this progress. still not sure yet but the 2.5GbE makes me think more about this one over the pi bpi-r4 as the only thing i see that one missing in the bpi-r4 is the 2.5GbE ports
I really believe in this project and following it. The only thing is other vendors like Unifi Dream Machine start at 350€ w/o tax. So compared to that, which features will this router config have and what not? Based on that, the estimate of 224-234 would be a max price I would consider unless it has a significant feature my current router does not have or it has optional features that make it to the 300 mark. But who knows, I am still following and curious about the final result and the unique selling point.
More memory might mean it can handle more advanced routing at or near linespeed. It does also have 2.5gbps ports instead of gbe for its secondaries, and you won't be locked into Ubiquiti's OS. Same here, I'm interested to see where this goes.
The way I see it there are essentially three groups of "customers": 1) The people who use their ISP-provided router or some Asus AC router with too many antennae - they probably won't see spending $300 on a router as worth it 2) Homelabbers - since this is basically a hobby for them you can probably get them to spend more (there's also the subset of open hardware dorks like myself who are willing to pay but will also want gerbers, etc) 3) Professionals/Businesses - They'll pay but you'll have to thread a needle to get to them, you might undercut the Traverse Ten64 (which you should look into), but as of right now I don't see the value prop that will convince the people who can afford Cisco gear to buy this instead Anyway, I'm rooting for this to become a real-deal open hardware product I can buy
I gave like 300USD would be considerable but I understand that is not realistic. But one of the questions that investors often ask is: what sets your product apart? I mean, I LOVE the design! I love the fact that it is crowd-sourced and that you share every step. But I am currently running a UDM Pro, which has 2x 10GbE, 9x 1GbE; 3.5 GbE IDS+IPS Throughput, has an HDD slot and NVR capabilities, has a SD-LAN software, comes from a reputable manufacturer (even if some wouldn’t consider Ubiquity as reputable) and has fun little features such as the AR view in the app. And that’s at around 450$ Unfortunately, for you, I am quite happy with my choice but maybe you could point out what makes your product better than all the competitors!
If it had a virtualization compatible intel CPU, so I could use it as a mini HomeLab or Travel-Lab, I would easily see this at 1000$ (upgradable RAM would be necessary).
4 2.5 RJ45 port and 2 SPF 10G. plus opensense / pfsense. that sounds like a router for me! If it even have a locking power connector it would be a steal for me
7:44 That’s where the “Economics of Scale” would join the scene 😅!! Now we need to validated the minimum quantity of parts that would make your BOM at least 10% less than best price based on charts and that’s from multiple vendors or negotiating each vendor to reduce at least 3%:5% and go to 10% if possible to accumulatively reach 10%25% BOM price reduction!
8:39 Also you could go second round users validation to assert of which features if removed they would find that useless? Compare this to their pricing practices they chose each and you could figure out new pricing opportunity by making different models delivering different features and values allocated based upon BOM cost and Final price opening a margin for you to leverage! Another idea is after making 2 different models at least based on user group’s analysis you could make a super premium version that have all features and extra add-one that’s low cost and convenient to justify to highest price point you’ve offered as a value priced model.
To be honest, without knowing what software do i have to use and how efficient / powerfull is, i dont know how much the router should cost. There are a lot of factors but just to mention a single one : Will it be capable to do VLAN-Routing at wire speed?
Your router just went from "not considering it for any price" to an interesting product. I just upgraded my internal network to 2.5 gigs and I need a better router for it. Also all my switches now support FSP+.
I appreciate your effort. i see the competition is tough. products popping up or coming to my attention, that are like what you try to build. for example a mini computer, driven by an intel N5000/N6000 Chip with 2 SFP+ Cages and 3 of 2,5GB i226V driven RJ45 Ports. there is even an version with an i3 N305 and a mellanox board for 2 * 25G SFP Ports, but at an pricepoint of 800Euro.
LOL, I just pulled $249 out of my bum based on previous videos! (My mental math before hearing this process -- 100 low end, 600 high end, ratio = 6:1, square root ~ 2.5, estimate sell price $249) 😃 Good to know my instincts are sound!
I am probably one of those that put too little in the price willing to pay category. In my case, as others have said, it's a matter of budget over value. Rather than answering what I think would be a reasonable price, I answered what I thought I could get approved for budget with my partner. We've been having some WiFi issues recently, that I think are partially resolved through using a better router at the start. I think before this product is released I would be able to justify a higher price point for a better router.
During the pause: Possible solutions to reduce the risk of going under before/after shipping: market segmentation (niches), multiple build options for different (and lesser or higher) market needs, and the correct prediction of the future prices of the parts going into the build - to allow increasing market share and reduced cost to drive gross & net profit. Also, unfortunately the reality is that shipping costs are a significant portion of the proposed price-profit, and *returns* will eat your lunch.
I absolutely love your solutions - thank you for chiming in! Based on my previous experience with DHL, a device like this shouldn't cost more than, say $20 to ship EU to US. I can't know for sure as I haven't sent one yet, but I'll probably do a "fake" router package (similar dimensions and weight), just so that I get a rough feeling. Thanks again! 🙌
I would try to make the case in a 3d printing farm, because you aren't tring to manufacture this product for mass production at the first stage, you will be able to mass produce in the future with companies like slant 3d will be able to manufacture at mass production, they can manufacture even on demand. and I think a router doesn't justify aluminum which is expensive to ship and manufacture. price: 350$ needs to be the price to make a profit (25% answered it's a great deal), even try downgrading to a 2.5Gb speed, it's even overkill for most people. another solution: you can make a monthly software that goes along and charge extra.
I'm just not buying the premise, you can install the operating systems you are talking about on most small single-board computers, all you need to do is get one with a dual port NIC, a "router" is only one piece of the puzzle, serious users will purchase their own switch, sorry but I just don't see your "product" being a success, all you have going for it is a nice case. The fundamentals are missing.
That formula in the beginning. only makes sense, in an non market - because the construct of a product is something that's already done - the formula your using more aligns with a blank sheet, and then building it, or adjusting it. Also, the will of the people, is the mapping of the opinions of the majority ;). To solve your "sensitivity mapping", the BOM/cogs/etc needs to be adjusted, and you need to decide a start margin, and a "end margin". There's no profit / no way of sustain the business with the production cost per unit with the information you put forward until 8:40 :). However, you can do batches, and target a smaller audience than the "sweetspot of the majority" - well, as long as you also take into account that you need to be able to sustain the business - or, take on additional capital and do the best you can to limit the "bleed" of money, and promote additional products - that you do this operation in reverse, thus, knowing if you hit a sweetspot or not :).
Getting conservative backers for this project I would gladly pay a little extra if it were expandable networks are always changing so if you could somehow future proof it say make sfp+ a daughter board to later be able to upgrade to QSFP+ or even be able to use 10GBe or 25GBe in the future I would pay more for it, but that might increase cost too. I'm still curious what market you are aiming this at cause I have a banana-pi r3 with dual 2.5gb sfp+ ports and because it only has 2 and 4 GBe ports that is why i'm looking to change. To compensate I'm using a 9 port 2.5GBe router with a sfp+ 10gb fiber port.
The problem is, that the chosen CPU can support 2x10Gb SFP+, but it cannot support any more than that, maybe a V2 or a "pro" model sometime down the line, but I don't want to get too much ahead of myself. As for the market, see my latest video ;)
Without reading comments nor continuing video beyond 8:40 my solution wold be: 1. Do a second survey asking people what features they consider useful and compare them with costs of those features. Remove what is expensive and not too useful. 2. Split final products into modular system. Remove from base version everything that has low value in #1 and add some of them back as skews or paid additions (if enough people want them). 3. Look for cheaper alternatives for components that are not critical for the product or have low results in survey from #1 edit: Now that I have seen the video, i think switching from 1GB to 2.5GB is interesting, but it will end up reducing number of ports from 4 to 3. I wonder if going 2x 2.5G + 2x 1G is possible. Also this could be one of the skews that I mentioned, where you could have the same device in 2 versions: either 4 port 1G or 3 port 2.5G. I am sure there will be market for both. edit2: If possible maybe even add a fifth 100mb port dedicated to connecting a printer or some other low traffic device. Low cost, added value.
I just checked the datasheet and you can have all the ports, so 2x sfp+ 2x 2.5G (yes 2) and 4x 1G, but I can't make out of they are all enabled at full speed see: "Layerscape LS1046A Reference Design Board"
No. For QSFP+ we'd need to pick a much much stronger CPU. One that's priced 3-4 times higher. The CPU alone. And it'd be a different class of device altogether. Plus, there isn't many home users (and SOHO) that need it. And I'm a firm believer to only add what 95% of people need 95% of the time.
The 300€ pricepoint from buyer side of view is perfect, the profit margin still needs to be acounted for. Sheetmetal case is profesional, bare PCB isnt realy. 3x 2.5G is awesome.
Good changes within the constraints, still sad no 25G option, 1-2U rack mount would also be a requirement for me, though, sadly, I have yet to find a router os/package I like as much as VMware NSX-T or Fortinets FortiOS.
25Gb won't be a thing in this router. It would at least double the price (given the CPU needs to be much faster) and honestly, home users aren't running 25Gb.
@@bigpod I WISH I had access to 25Gbps WAN, I am trapped on DOCSIS but my house is only ~100 feet from one of the areas huge fiber trunks. What I need the 25G+ ports for on a router are L3+ routing or traffic filtering/inspection between VLANs. This is still my house, I don't want to have to deploy bare metal edges for NSX-T and setup BGP and put up with the supper buggy UI in NSX-T and some of it's other issues that make me wonder why they retired NSX-V when they did and shipped/pushed people to T way before it is production ready.
I'll take 10! Using a BPI router board now. BPI is ok but it's a race to the bottom. No upgradeable wifi and limited pcie/m.2 options. Plus no good case options.
In that case, you're out of luck when it comes to this device. We're making a router, and because the CPU has already been settled on, there's no way for it to handle 4 SFP+ ports. And even if we chose one that supports what you want, then the device would get more way more expensive. And you see where the sweet spot is.
@@tomazzaman if it is both then there is no problem. But there is a very high probability of people using it without the enclosure leading to early deaths of the device due to dust/cat attacks etc.
N100 has no embedded options available. Meaning Intel doesn't guarantee any amount of time of production. So it might get deprecated a few years from now, going against what this router is all about (longevity). Plus, as you mention, doesn't have any 10Gb networking interfaces, so we'd need to add those chips, likely getting close or over what the NXP CPU costs. And finally, likely most importantly, NXP CPU is highly optimized for networking. Meaning it'll be faster and way more efficient at networking tasks than a general purpose N100.
CWWK and their ilk (Topton, Minisforum, Beelink, etc) have the "do-everything miniPC" niche pretty well sewn up. A genuinely 10gig-capable router+firewall more or less doesn't exist sub-enterprise-tier, so he's got a lot better opportunity there imo. Hell, *I* couldn't justify it currently, but even at $750 he'd be firmly the cheapest thing in his performance/capability segment- and if he can swing a solid VPP implementation to open up future upgrade tiers with 25 or even 40gbps nics without needing dedicated ASICS or massively overpowered general-purpose CPUs, well, the sky's the limit. Granted, those are a wide selection of medium-to-large-sized "ifs". But I'm real interested to see how it turns out for him.
From having been in startups and a few product stints, beware of survivorship bias in your surveys. The problem with relying purely on followers and subscribers is their inbound marketing nature. You won't get feedback from the demographic that might actually be the true power users and even influencers of their communities, leaving you without feedback that you need to steer your project correctly.
Yup, agreed! I've done a lot of 1-on-1 research with some people from my personal network and the numbers were significantly higher. Last phone call I had, the other person literally said: "Price it 600, and I'll take two!" :)
@@tomazzaman+1 to that. Take a look at some stats for current high end routers, especially on enterprise products. Back when I still ran game and email servers, the challenge of keeping a router up was ventilation. If I wanted 24/7 cooling for a router with SoC churning at 15W, I had to get a server 1U product at least, which costs significantly more. It's worth spending some time to look through current offerings from various places like Fortinet, Palo Alto, Dell, etc.
You really should focus on the router part and the advance home user market. Personally I will have a separate switch anyway (I will need more than 3-4 ports), but what I was looking for is just a reasonably priced 10GB router with at least 2 10GB Base-T RJ45 Ports (WAN/LAN) because most Fiber-Modems have a RJ45-Out and there are already decent priced RJ45-10GB-Switches. But there are no “cheap” 10 GB routers for the home market. And RJ45 makes it all easier for the (advanced) home user. Why SFP+ ports? You shouldn’t try to compete with professional products. If SFP+ ports, at least make 2 & 2 RJ45/SFP+ combo ports (2 WAN/2 LAN)
I hear you loud and clear, but I don't agree with the last bit - I shouldn't compete with professional products? Why not? Because I definitely will. This router will be as professional as they come.
Honestly, I don’t think you will have any chance to enter the professional market. No offense. But this is just another game. But bringing value to the advanced homelab users with 10GB Fiber connection could be your niche.
10G for the advanced homelab users is my niche. But that doesn't mean the product won't be professional. No offense taken, I think it's the language barrier.
If you can bring a professional 10GB router to the homelab community for about +/- 500€ (or maybe even less) I am all in. But please with at least 2 10GB RJ45 ports ;-)
I would employ another youtuber they watch to give the product a favorable review. Regarding the cost, I'd find and butcher some existing cheap piece that is close enough to use as a starting point. If it's not broken, don't fix it.
Nisem sledil čisto temu projektu ampak, po mojem Wifi-ja se ne rabi. Samo ustreliš se v nogo, ker ljudje potem pričakujejo nek hud performance, ki ga ni in je slaba volja. Kaj pa kakšn extruded aluminium? IMO je storage lahko 32GB Storage bi lahko bil samo M.2 slot Nabavne cene moraš po mojem še dodatno dol zbit, razmišlat o kakšnem Kickstarterju in si zadat višji cilj kot 1k enot. Ram je predrag. A 4GB ne bi bilo dovolj? Kaj se pričakuje, da bo uporabnik naložil na to zadevo? Nek OPNsense, PFsense? Po mojem je 4GB dovolj. Naj bo več rama za verzijo 2.0 al pa Pro verzijo :) Sicer pa kul projekt. Upam, da ti uspe :) PS: Osebno sploh ne bi SFPjev dajal, ker 10Gbps je mal huda za to zadevo. Za neke dolge razdalje tud najbrž ni namenjen, da bi optiko fural. Za se izognit ISPjevi opremi pa rajši switch uporabiš, ker imaš najbrž še multicast, ki ga nočeš čez to zadevo pošiljat. Oziroma, če že mora bit, bi dal kot optional addon, ki si ga lahko sam vgradi uporabnik. Tudi marža je višja na taki zadevi potem. Pa lahko pol dodaš še kakšen M.2 extension al pa podobno, lahko tud kakšen SATA controller, če bi kdo hotu naredit potem kakšne vrste NAS iz tega. LP, Tomaž (Tudi tukaj) ;)
Wifi-ja v originalno ne bo, bo pa mini-PCIe port, da se ga lahko doda kasneje. Extruded alu je grd, vsaj za moj okus. Sheet metal bo. Kar se pa portov tice, pa je precej koncen spec (2x SFP+ in 3x 2.5G RJ-45), ker smo z izbiro procesorja zakljucili. Gor bo najprej OpenWRT, ker je najbolj podprt, kasneje upam, da uspe nalozit OPNsense.
Apple proved that this method is no longer valid, you can continually lower the quality, limit what the customer can do with the device while increasing the price!
I appreciate the feedback nonetheless. And the good part about SFP+ cages is that you can actually put in an RJ-45 transciever, turning it into a copper connection! 💪
@@tomazzaman yea but it is an added cost but that makes sense. I hope this goes Weill for u. I found your channel from the isp router video that just blew up
Xiaomi is expanding its lineup of WiFi 7 routers with the new BE 3600, set to launch on January 30th in China. While the official price remains under wraps, pre-reservations are now open which shows the price will not be more than 299 yuan (around $43). very professional to make something that costs 10x more . applause!
The problem, I think, with this method is that people have different incomes/budgets and "expensive" is relative/subjective heh And people wanting to pay less than the hardware actually costs is a problem with capitalism heh This is exactly why I will never become an inventor heh
With a nearly 500$ price, and removing the original case, you are now in the price point of a dreammachine special edition from ubiquiti. Wich have integration with software and harware ecosystem, and have a descent support and waranty. I like your project, i will enjoy watching your video, but I will not buy this product.
Lets be honest anybody willing to pay less than 600 has no idea how expensive any 10Gbit purpose built firewall is. I mean look at netgate or opensense hardware, bare minumum for a 1-2 port 10Gbit FW is 800+ And stuff from "proper" brands is at least 1k, usually significantly more than 1k. And if you want to build your own FW, and don't want to go with 5+ year old, loud/ less efficient hardware you'll have still have it hard to stay below 600.
Yep, I don’t have much to add. I’m doing my best to optimize every single bit to bring the costs down but still, 10Gb requires some muscle, and muscle doesn’t come cheap.
Here is what i already wrote on previous iteration. Im one of those who said way to little but that is mainly because of im without a job(used to work as a programmer) so i value everything as gorenc ;). you will know my idiotic anwser by my email as i have custom domain (same as my youtube username) when i remove my current situation out of the equasion i can say that i would be willing to pay 500+€ for it but less than mac mini. if you want to ask questions im available(clearly since i do not have a job) and you already have my email :) so just send the questions my way PS. as a fellow youtuber im rewatching your video for algorithms sake
From an accidental search regarding fiber you are overall becoming an enjoyable TH-camr, with passion, entertainment, and your entrepreneurial spirit is inspiring me to do my own thing. I’m so excited for your router. I’ll probably be one of the first. I’m ready to replace my ubiquiti gear.
Thank you! Comments like these add so much fuel to my fire I can't even begin to explain! 🙌
@@tomazzaman Keep it up! It’s impressive watching your growth, even the last few weeks.
Your videos feel like you’ve been producing for years. You also encapsulate entire topics to clear undertake parts, and English isn’t even your first language! Have a great 2024! Excited to see you take off! We’ll all be here for the ride!
Other than the router what do you want to replace?
I currently own a service at my company and have dealt with this same issue. Typically more often than not, the customers who think the product is too expensive don't understand the value of what they are receiving. In my case, I usually ensure they understand the service and compare it to competitive services as well as what it would cost them to hire someone to provide themselves with the service.
Thanks for the great video! Keep up the good work!
Yep, it can be quite frustrating to teach people of the value we provide. Think of it as a good challenge! :) Thanks!
From hacking his ISP's router to this. Love the content. My original answers were made before watching the BOM video, and seeing it I kinda wanted to revise my answers seeing how much it was likely to cost. I believe $300 plus whatever the margin would be seems fair honestly.
Thanks! Appreciate the feedback, a couple of other people said the same thing, so I know I'm on the right track!
One thing at is (yet) missing from that market research is to determine what would be the target comparable existing products and what is their average sale price. Comparison should look at features, performance and warranty. For example, if all comparable products cost a little more than 500$, then people looking for that product will still be attracted to it at 500$ because it is still cheaper than all the alternatives in order to get the same value.
In order to determine the comparable products, a performance benchmark of the total throughput should help.
In my opinion, people who only expect to pay 250$ for a router want extra features that they probably would not use, but still say "it's a nice to have" for sometime in the future without having an actual plan to start using them. People with an actual "business" need for the extra feature are the one that know how much that feature is worth. I wonder how many people from that survey do not even have a 10G capable network and have no plan on upgrading...
On a purely business point of view, you would probably make as much profit by serving the top 20% higher responders than the mass mid-range.
It's less about the features and more about how much pain and what types of it the different market segments are willing to accept. One segment may be willing to spend thousands of dollars on a 10gig router while another is only willing to do a few hundred. Why such a difference in price for the same feature? it comes down to what pain points they experience the most. Examples:
*Government:* Major pain points are organizational. It's a PITA to get any new vendor authorized to work within a governmental agency. Willing to blow a lot of money to not have to deal with their own bureaucracy. Normally siloed into a few vendors.
*Large Enterprise:* Similar to government, but may focus more on "scale-out" and remote management/automation features. Similar pain points but with a bit more wiggle room. Still won't deal with small vendors very often.
*Small-Medium Enterprise:* Pain points start to focus on cost/benefit relationship of products used. Similar requirements to large enterprise needs for management and uniformity, but more often willing to take risks on unknown products in order to get a better TCO.
*Small business:* Pain point is money and uptime. Quite often they'll be willing to cut back on features in order to find something stable and consistent. Quite often they also don't have many manhours available to focus on technology and instead need higher usability requirements compared to SME.
*Owner/Operator Business:* Money and uptime, but to lesser degrees. Generally they want to use off-the-shelf parts that they have to think about once and never again. Since they don't have many/any employees but themselves they'll be willing to put up with downtime as long as the cost is low enough. Also quite often they'll hire someone else to set up their equipment and actually don't make the purchasing decisions themselves. The people they hire likely have pain points similar to SME or Geeks, depending on who they hire.
*Geeks:* People who love tech, quite often spend large amount of money on products as status symbols or because it's the new thing. Early adopters, their pain points are individual but ease of use and product appearance needs are high for much of the market segment.
*Homelabbers:* Geeks but they operate more like an SME. They want high-end equipment for cheap running large feature subsets since they want to learn about technology in-depth. Quite often willing to put up with large amounts of technical and stability pain as long as the product is performant. Primary pain point is usually cost (even though they may spend thousands more than a "Geek" would).
As it stands, the segments this new router is best oriented towards are the small business and geek segments. Maybe SME and homelabbers, however:
-As a homelabber I'd be aiming at something like the CCR2004-1G-12S+2XS from Mikrotik at the $500 price point, or purchase used enterprise equipment.
-As an SME I'd be aiming at something from second-tier companies (Ubiquiti, Ruckus, Aruba, Dell), Juniper or Cisco. All of those products have some form of remote or cloud management service and offer a lot of capability and stability at a decent price. They also provide uniformity for the network with brands likely to still be selling products in 10-20 years' time.
Just found this product so I'll need to look into this router more indepth before I say whether I'd buy it or not. But, that's my general thoughts.
@@jasonmonsen7052 I love you answer.
This new router being essentially a piece of hardware, it does not do anything by itself. It needs a firewall software running on it to create its value. If the targeted softwares are OPNSense, PFSense and VyOS, then the implicit market segment for the hardware will be the same ones as the software that will be running on it, or maybe a subset of them.
This is probably what was missing in the survey that gave such a wide range of expected sale price between the responders: knowing what market each responder were identifying themselves to. The geeks or owner/operator were surely not answering the same way as someone responding with a large business in mind.
@@viaujoc Your first couple sentences leads into what I was thinking about yesterday after I posted. Tomaz' follow up video about adding value will likely cover this, but one of the things that could set this product apart is a dead-simple out-of-the-box experience for the router itself. I don't know how hard it'd be to create some wizards within the mentioned OSes, but if you could offer the raw power while also offering an experience similar to an off-the-shelf Linksys box, I think it'd be a winner. Some unique value-adding features to the software could be a strong selling point.
Thinking about what group an easy to use powerhouse would appeal to, I think of: gamers, streamers, and any small company working with CAD, video, or AI systems. Dual 10gip with 4 2.5gbps ports would be pretty nice for anyone working with large data in a hybrid setup (cloud and on-premises), or small offices with off-premises cloud infrastructure. Tomaz being a YTer likely has some good ideas around what advanced features would be useful for streamers. Things like QoS or VLANs could be very useful to creators who could benefit from it but might not have enough experience to know what they're doing. Including some newbie-friendly guides that go over some creator-centric features which might be useful to them (and giving examples of when or why) and step-by-step instructions (or wizards) to add these features would help a lot.
For more technical "geeks/homelabbers" that value-add might not be as useful, but it wouldn't detract from the product itself as long as there's an advanced mode. I did mention a good Mikrotik alternative at near the same price, however it is a weaker processor and only has 12x1gip ports alongside its 2x10gip, so this one could make sense in certain network configurations. For gamers, while this switch would be overkill, 2.5gip NICs are starting to show up on motherboards and I could see myself justifying the purchase. As another commenter mentioned on this video, most "gaming" routers are complete crap and a money grab, offering a legitimate "prosumer" option would be something fairly unique.
I wouldn't focus on selling to businesses at first with the version 1 (outside streamers and some large dataset studios). If this does see multiple revisions and becomes something longer-term then I could see businesses considering it. The port layout is pretty great for a flashable router and would beat out a lot of the x86 whitebox competition.
@@viaujoc Just a heads up your last post is hidden for some reason. And my last answer was kind of talking "at" you instead of "to" you. You're right though that it makes a lot of sense to understand how people were approaching their dollar values. I don't think many would be coming at this from a large business point of view (I already have vendor contracts), but it is still very interesting to know what they think their use case for such a router would be.
@@jasonmonsen7052 Unfortunately integrating third party open source softwares is not an easy feat. Some softwares (ex: pfSense CE) have explicit restrictions in their licence agreement prohibiting selling a hardware with their software pre-installed. However, additional value may be created by giving the buyer some guaranty that a specific third party software product runs stably on it by thoroughly testing it and providing installation guidelines and best practices to integrate the hardware and software together. PCEngines did that and had a respectable product with a good sales volume for many years.
I also agree with you that targeting medium to large businesses on day 1 may not be a good idea because those companies will usually be looking for maintenance contracts and strict response time commitments. This involves having a dedicated support team with the salaries and other related expenses.
Nice discussion we have here. I hope this will be useful to Tomaz...
One other thing I'd like to mention, power efficiency. Most of the time hardware is idling, so if the idle power consumption can be optimised it would help with those of us paying crazy amounts for power.
Aye. The CPU we chose is a very efficient one, single digit watts while idle. Once we get the development kit, I’ll make a dedicated video just about this topic.
I would try to make the router look more expensive w/o actually increasing it's cost (much). I also wouldn't just show them a router, I would explain its benefits over the competition.
Valid points! 🙌
Tomaž, you have officially become my favourite Slovenian youtuber :) Maš to!
Hvala!
love this series. thanks for sharing
stamped anodized aluminum sheet metal would be a good compromise
glad to see others agreed with a cheaper case and the addition of 2.5gbe ports, look forward to seeing the end product
Thank you!
Agreed. I do have a slight preference for aluminum over plastic, but formed sheet is fine. CNC hogouts seems like gross overkill. And, I have plenty of plastic stuff too, so even that's not an insurmountable barrier. I do industrial controls, so while I'd like one for my own use, to build one into equipment I sell, I'd like DIN rail mount in bookshelf orientation, actual terminals for power vs. a barrel connector (preferably 24VDC) , and all the lights and ports on the exposed edge.
Excellent! Perceived value was one of the tools British Airways used to price up the cost of tickets for Concorde flights to the US - finally turning a profit after years of losses by *increasing* the ticket prices!
This video made me go back and check your other videos about your router plans. I'll be watching this with interest. I recently watched a video on "Serve the Home" channel, the video is titled "The EVERYTHING $300 Fanless Home Server". Since you're figures were around the same I would suggest looking at that hardware and maybe thinking about what you can copy in terms of BOM and also differentiate in terms of features.
Thanks for the suggestion, I’m a big fan of Serve The Home!
If you're thinking about shipping globally, have you considered asking people about regional price sensitivity? Different countries have differently-scaled economies after all, and that will affect how much people can afford to pay for anything.
Good question! I have not, because this will have to do for now, but will definitely in the future!
Amazing. I actually learned alot. Thanks for documenting your journey and I wish you luck. You might end up making my a backer at the end of this!
Thank you, will do my best!
Awesome video Tomaž! It is very helpful for the ones of us, who have little idea what goes into the pricing parts of product development! 😄 Ži čakam na nasledni videjo
Hvala! :)
If you continue like this, there is great future ahed of you!
Ngl i imagined from the get go that a Premium version of the product would be obvious.
loving the idea and the drive to get a good open-source router on to the market! I put my own router together using a Lenovo mini PC, as I couldn't find what I wanted on the market. One nice to have feature, if I was looking at building a router from scratch, a dedicated management port that can accept POE+ to power the unit. This allows the provision of redundant power if needed, or just makes an install much cleaner :) That said, your design ticks all the wish list items on what I would want on a router, multiple SFP+, 2.5GB LAN, good IDS/IPS performance, a wide choice of OS. Keep up the awesome work!
i would definitely love to see a video comparing your product to the competition. maybe an average isp modem vs an average mid range router vs an average high end one on the market right now.
My solution would be to offer the router in variations that cut down on some of the features less useful for an average user - like having 64 GB of onboard storage... In a router?.. And lower the price point for that variation accordingly.
That way, you could get both the customers who believe the current iteration of your router is too expensive (likely because they won't be making use of a lot of its features) and the ones who are willing to pay what it's actually worth (likely the people who appreciate all of its abilities in full).
For the challenge- I would ask myself whether the respondents to the survey really are the ideal customer I’d want to sell to. If not, I’d run the survey again and try to get a lot more targeted about who I try to get the survey in front of. If so, I’d do a few different things: see if I could reduce COGS, and since that’s unlikely without significantly sacrificing quality, make sure I thoroughly educate the potential customer before presenting the price. Another idea could be to offer a lower quality product (1G networking or maybe 2.5?) for the people who want a lower price point. Could even skip the aluminum case on that one, or use an aluminum case that’s industrially available for mini-pcs and possibly cut your costs that way.
Considering the fact that some of the Protectli Vaults go for $500-$700, I think this is excellent value for what it is, especially with 10G.
Hi Tomaz...a great video! You have gotten us to think about how things are made and priced. Looking forward to seeing where this goes. I think you are on the right track!
Thanks! Glad to be of service! 🫡
Make a tiered router= make a pcb that if fully populated has every option ( al the GB/2.5G & 2x SFP+ caged) and a nice anodized alu case (400-500€ range). When only populated with 1x sfp+ & 1x 2.5G & 1x GB port and a sheet steel housing then it's the 200€ range (minimum edition). You can do a H4X0r edition to: here's the pcb and components have fun ;) or a Lite editon = same as minimum but no case.
I see you had the same idea as me, and also 2 weeks late to watch the video :)
@@hubertnnnYeah so many reactions, but it's always good to see other people with the same idea :D
I'm surprised how many have unrealistic expectations to what a product like this cost. I'm guessing they are either not very well versed in the networking world or are from countries with a much lower per capita income than the average.
I consider this product to be a prosumer product or SOHO if you will. And based on the current specs, compared to similar products, I think the "founders edition" fits nicely somewhere in the 500-600 USD range, while the cheaper one (plastic instead of aluminium) in the 500-550 USD range.
I have a router from Netgate, so I'm not looking to replace it yet, but it's interesting to follow this journey. It's both entertaining and educational.
The quality, both in terms of content and presentation, of this channel is awesome, keep up the impressive work Tomaž :D
Thank you! And honestly, I have nothing to add, I pretty much agree with everything you just wrote.
Its most likely regional based. I mentioned this in a comment in previous video.
For example in Poland which is a poor country, the most similar to this router devices are:
Ubiquiti X-SFP for 429 PLN (106 USD)
QNAP QSW-2104-2S for 629 PLN (156 USD)
I initially searched different stuff and watched your other videos, but you got my interest with announcements of building your own router I wish you good luck, you got me interested in end product. One thing I seams to miss is what is your intention for target mass of buyers, business/home or ??? Especially as I am trying to figure out what is your answer on the same 4 questions you wanted from your subscribers in order to get price point.
Hope you will clear few of these questions in next videos, because target usage would be one of the turning point for some target buyers, me included. eg. ground up design of your LAN/WLAN or you are searching for expansion or step to evolve your existing infrastructure which greatly influence how much you pay for a device and if it will fill your purpose.
Would be glad to talk about it with you and your subscribers. Hope will see replies.
Thanks for the kind words! I'm currently focusing on the advanced home users because I'm one of them, but do have plans to address a couple of other markets in the future! I do plan to make a video about that as well, probably not just yet, but when things really clear out.
@@tomazzaman That is the answer I wanted to hear, because like you I am advanced home user, very interested in your product which from my brief analyze fits exactly my needs for mid range network infrastructure complexity and future proof that would allow me to archive my home plan and level of independency and control over it.
Hvala, veoma uživam u tvojim videima.
What OS are you planning to use? Will the user be able to change? Are you sticking with unchangeable soldered RAM & Storage?
Awesome stuff. I am not a techy person at all but i like seeing tech videos. Nicely done.
Thank you!
It also sounds like you can use the price sensitivity data to produce a graph of "expected profit" by taking in the raw cost, subtracting the price point, and multiplying by expected percentage of clients. Then, use that to maximize profit (without going to the extremes, of course). I wonder if that's a reasonable approach.
It is, but the PSM gives you more of a general idea than a final one. For example, imagine you produce the unit overseas. This means you also have to ship it to the warehouse from which you sell. This also adds costs, that are not in BOM. And there’s plenty of costs like this that are hard to account for this far in advance.
@@tomazzaman I see! Thanks for the explanation
After a long doom scroll i came across you, been hooked since. Will buy the second it hits the market, i need to get rid of the prop HW our ISPs in India give.
You could try SLS 3d printing for the case. It can 3d print both plastic and metals!
It would be an idea you could try.
Have you considered crowd sourcing some models that can be 3D printed at home for the case and selling the the rest in a kit? Not to mention people buying “high performance routers” likely want a way to easily mount it in their network rack. Any thoughts?
A highly intriguing project & amazingly entertaining videos. And for me, watching every next video is also adding to an increase of the perceived value and/or willingness to maybe pay more for the unit. (I previously filled in the Pricing form)
Ha! It IS WORKING indeed! :) Thanks for the kind words!
$649, maybe $699 at it's current COGS/BOM. A $549/$599 price point would be ideal if you could get the costs down though.
Yep, agreed. Doing my best to bring them down.
Good luck. Lot of money for normal people and us nerds know hoe to buy better kit for less$ small network appliances is a pretty saturated marketplace, good luck beating some of cwwk's latest.
Great video. In the next video could you discuss what operating systems we'll be able to run on this ARM in more specific detail? For example, will there actually be an ARM build of OPNsense built and made available? - I know pfSense has an ARM build but it's currently restricted to their own hardware appliances and I don't believe OPNsense has an ARM build yet.
I have it planned, but not in the next one, I have a couple of other videos to make before we get to the actual software and hardware, primarily because I don't have the development kit yet and also because I want to share the full journey and not skip any steps. Thanks!
@@tomazzaman Thank you, I look forward to all those videos :)
$159, my final offer...
No.
id like to think i had a hand in changeing out the one 10GbE to another SFP+ Port as i suggested, and i will continue to watch as this progress. still not sure yet but the 2.5GbE makes me think more about this one over the pi bpi-r4 as the only thing i see that one missing in the bpi-r4 is the 2.5GbE ports
I'll take it! Just give me a couple of more videos and I'll surely convince you! 💪😅
I really believe in this project and following it. The only thing is other vendors like Unifi Dream Machine start at 350€ w/o tax.
So compared to that, which features will this router config have and what not?
Based on that, the estimate of 224-234 would be a max price I would consider unless it has a significant feature my current router does not have or it has optional features that make it to the 300 mark.
But who knows, I am still following and curious about the final result and the unique selling point.
More memory might mean it can handle more advanced routing at or near linespeed. It does also have 2.5gbps ports instead of gbe for its secondaries, and you won't be locked into Ubiquiti's OS. Same here, I'm interested to see where this goes.
The way I see it there are essentially three groups of "customers":
1) The people who use their ISP-provided router or some Asus AC router with too many antennae - they probably won't see spending $300 on a router as worth it
2) Homelabbers - since this is basically a hobby for them you can probably get them to spend more (there's also the subset of open hardware dorks like myself who are willing to pay but will also want gerbers, etc)
3) Professionals/Businesses - They'll pay but you'll have to thread a needle to get to them, you might undercut the Traverse Ten64 (which you should look into), but as of right now I don't see the value prop that will convince the people who can afford Cisco gear to buy this instead
Anyway, I'm rooting for this to become a real-deal open hardware product I can buy
Thank you! You're right. I have a couple of ideas about the 1) and 2) 💪
I gave like 300USD would be considerable but I understand that is not realistic. But one of the questions that investors often ask is: what sets your product apart?
I mean, I LOVE the design! I love the fact that it is crowd-sourced and that you share every step. But I am currently running a UDM Pro, which has 2x 10GbE, 9x 1GbE; 3.5 GbE IDS+IPS Throughput, has an HDD slot and NVR capabilities, has a SD-LAN software, comes from a reputable manufacturer (even if some wouldn’t consider Ubiquity as reputable) and has fun little features such as the AR view in the app.
And that’s at around 450$
Unfortunately, for you, I am quite happy with my choice but maybe you could point out what makes your product better than all the competitors!
If it had a virtualization compatible intel CPU, so I could use it as a mini HomeLab or Travel-Lab, I would easily see this at 1000$ (upgradable RAM would be necessary).
@@zanderfaehrteisenbah You could buy something like the Qotom Q20332G9 or the Minisforum MS-01.
I take 3NIC 2.5Gb/s instead of 4NIC 1Gb/s all day.
Yep. Same! :D
4 2.5 RJ45 port and 2 SPF 10G. plus opensense / pfsense. that sounds like a router for me!
If it even have a locking power connector it would be a steal for me
7:44 That’s where the “Economics of Scale” would join the scene 😅!!
Now we need to validated the minimum quantity of parts that would make your BOM at least 10% less than best price based on charts and that’s from multiple vendors or negotiating each vendor to reduce at least 3%:5% and go to 10% if possible to accumulatively reach 10%25% BOM price reduction!
8:39 Also you could go second round users validation to assert of which features if removed they would find that useless?
Compare this to their pricing practices they chose each and you could figure out new pricing opportunity by making different models delivering different features and values allocated based upon BOM cost and Final price opening a margin for you to leverage!
Another idea is after making 2 different models at least based on user group’s analysis you could make a super premium version that have all features and extra add-one that’s low cost and convenient to justify to highest price point you’ve offered as a value priced model.
Great project presented enthusiastically and well structured. Subscription earned 😊.
Can't help myself 😂
To be honest, without knowing what software do i have to use and how efficient / powerfull is, i dont know how much the router should cost. There are a lot of factors but just to mention a single one : Will it be capable to do VLAN-Routing at wire speed?
Your router just went from "not considering it for any price" to an interesting product. I just upgraded my internal network to 2.5 gigs and I need a better router for it. Also all my switches now support FSP+.
I appreciate your effort. i see the competition is tough.
products popping up or coming to my attention, that are like what you try to build. for example a mini computer, driven by an intel N5000/N6000 Chip with 2 SFP+ Cages and 3 of 2,5GB i226V driven RJ45 Ports. there is even an version with an i3 N305 and a mellanox board for 2 * 25G SFP Ports, but at an pricepoint of 800Euro.
LOL, I just pulled $249 out of my bum based on previous videos! (My mental math before hearing this process -- 100 low end, 600 high end, ratio = 6:1, square root ~ 2.5, estimate sell price $249) 😃 Good to know my instincts are sound!
Awesome video. How about maybe 3D printing the case so it's cheaper? or just giving people the file to 3D prints, so it's even cheaper?
Yes, we plan to release the STEP file of the PCB once it's final. And not only that. We plan a competition who'll make the best looking case!
wow that is incredible. I'm not even that much Internet networking, and I still want it.@@tomazzaman
I am probably one of those that put too little in the price willing to pay category. In my case, as others have said, it's a matter of budget over value. Rather than answering what I think would be a reasonable price, I answered what I thought I could get approved for budget with my partner. We've been having some WiFi issues recently, that I think are partially resolved through using a better router at the start. I think before this product is released I would be able to justify a higher price point for a better router.
Great initiative, but how about software production cost? I would consider openwrt capability if possible.
OpenWRT is the first piece of software we aim to deploy because of its support for a wide array of hardware. Next step: OPNsense.
@@tomazzaman Good plan 👍
During the pause: Possible solutions to reduce the risk of going under before/after shipping: market segmentation (niches), multiple build options for different (and lesser or higher) market needs, and the correct prediction of the future prices of the parts going into the build - to allow increasing market share and reduced cost to drive gross & net profit.
Also, unfortunately the reality is that shipping costs are a significant portion of the proposed price-profit, and *returns* will eat your lunch.
(and again, I have no idea why YT shows some random string for my username.)
I absolutely love your solutions - thank you for chiming in! Based on my previous experience with DHL, a device like this shouldn't cost more than, say $20 to ship EU to US. I can't know for sure as I haven't sent one yet, but I'll probably do a "fake" router package (similar dimensions and weight), just so that I get a rough feeling. Thanks again! 🙌
I would try to make the case in a 3d printing farm, because you aren't tring to manufacture this product for mass production at the first stage, you will be able to mass produce in the future with companies like slant 3d will be able to manufacture at mass production, they can manufacture even on demand. and I think a router doesn't justify aluminum which is expensive to ship and manufacture.
price: 350$ needs to be the price to make a profit (25% answered it's a great deal), even try downgrading to a 2.5Gb speed, it's even overkill for most people.
another solution: you can make a monthly software that goes along and charge extra.
Just FYI, the "too cheap" option didn't go below 100 dollars.
maybe also add some mini UPS inside at 15W you can get 1-2h of use easy (tho an external power bank would also do they trick.
at this point won't it be cheaper to use n95/n100 instead of an arm chip?
I'm just not buying the premise, you can install the operating systems you are talking about on most small single-board computers, all you need to do is get one with a dual port NIC, a "router" is only one piece of the puzzle, serious users will purchase their own switch, sorry but I just don't see your "product" being a success, all you have going for it is a nice case. The fundamentals are missing.
That formula in the beginning. only makes sense, in an non market - because the construct of a product is something that's already done - the formula your using more aligns with a blank sheet, and then building it, or adjusting it. Also, the will of the people, is the mapping of the opinions of the majority ;).
To solve your "sensitivity mapping", the BOM/cogs/etc needs to be adjusted, and you need to decide a start margin, and a "end margin".
There's no profit / no way of sustain the business with the production cost per unit with the information you put forward until 8:40 :). However, you can do batches, and target a smaller audience than the "sweetspot of the majority" - well, as long as you also take into account that you need to be able to sustain the business - or, take on additional capital and do the best you can to limit the "bleed" of money, and promote additional products - that you do this operation in reverse, thus, knowing if you hit a sweetspot or not :).
For that price there's quite good mikrotiks
Getting conservative backers for this project I would gladly pay a little extra if it were expandable networks are always changing so if you could somehow future proof it say make sfp+ a daughter board to later be able to upgrade to QSFP+ or even be able to use 10GBe or 25GBe in the future I would pay more for it, but that might increase cost too. I'm still curious what market you are aiming this at cause I have a banana-pi r3 with dual 2.5gb sfp+ ports and because it only has 2 and 4 GBe ports that is why i'm looking to change. To compensate I'm using a 9 port 2.5GBe router with a sfp+ 10gb fiber port.
The problem is, that the chosen CPU can support 2x10Gb SFP+, but it cannot support any more than that, maybe a V2 or a "pro" model sometime down the line, but I don't want to get too much ahead of myself. As for the market, see my latest video ;)
Thanks!
Thank you! ❤️
Without reading comments nor continuing video beyond 8:40 my solution wold be:
1. Do a second survey asking people what features they consider useful and compare them with costs of those features. Remove what is expensive and not too useful.
2. Split final products into modular system. Remove from base version everything that has low value in #1 and add some of them back as skews or paid additions (if enough people want them).
3. Look for cheaper alternatives for components that are not critical for the product or have low results in survey from #1
edit:
Now that I have seen the video, i think switching from 1GB to 2.5GB is interesting, but it will end up reducing number of ports from 4 to 3.
I wonder if going 2x 2.5G + 2x 1G is possible.
Also this could be one of the skews that I mentioned, where you could have the same device in 2 versions: either 4 port 1G or 3 port 2.5G.
I am sure there will be market for both.
edit2:
If possible maybe even add a fifth 100mb port dedicated to connecting a printer or some other low traffic device.
Low cost, added value.
I just checked the datasheet and you can have all the ports, so 2x sfp+ 2x 2.5G (yes 2) and 4x 1G, but I can't make out of they are all enabled at full speed see: "Layerscape LS1046A Reference Design Board"
3x2.5Gb OR 4x1Gb. Not both. And I'm willing to lose a port if that means the rest are of a higher speed.
@@tomazzamanI hate datasheets for that part :(
Question: would a swith to QSFP+ ports (instead of adding more SFP+) satisfy people that are port-hungry, while keeping the price of goods acceptable?
No. For QSFP+ we'd need to pick a much much stronger CPU. One that's priced 3-4 times higher. The CPU alone. And it'd be a different class of device altogether. Plus, there isn't many home users (and SOHO) that need it. And I'm a firm believer to only add what 95% of people need 95% of the time.
The 300€ pricepoint from buyer side of view is perfect, the profit margin still needs to be acounted for. Sheetmetal case is profesional, bare PCB isnt realy. 3x 2.5G is awesome.
Good changes within the constraints, still sad no 25G option, 1-2U rack mount would also be a requirement for me, though, sadly, I have yet to find a router os/package I like as much as VMware NSX-T or Fortinets FortiOS.
25Gb won't be a thing in this router. It would at least double the price (given the CPU needs to be much faster) and honestly, home users aren't running 25Gb.
25G for a router is way too much most SOHO and home users dont have a 25G pipe to internet
@@tomazzaman I mean, I am mostly 25 or 100G only falling down to 1/10G for things like IPMI or cameras, or, sadly, access to the internet.
@@bigpod I WISH I had access to 25Gbps WAN, I am trapped on DOCSIS but my house is only ~100 feet from one of the areas huge fiber trunks. What I need the 25G+ ports for on a router are L3+ routing or traffic filtering/inspection between VLANs. This is still my house, I don't want to have to deploy bare metal edges for NSX-T and setup BGP and put up with the supper buggy UI in NSX-T and some of it's other issues that make me wonder why they retired NSX-V when they did and shipped/pushed people to T way before it is production ready.
@@gcs8 then maybe enterprise offerings are for you
I think you have to expand your UPSses in contrast to other players on the market, like Gowin 25gbe
I'll take 10! Using a BPI router board now. BPI is ok but it's a race to the bottom. No upgradeable wifi and limited pcie/m.2 options. Plus no good case options.
*writes down to make 10 extra units. 😂
i want 4 sfp+ ports
i want it to be an intel x64
and
and
i should be able to install proxmox on it
as right now there is nothing much on the interwebs
In that case, you're out of luck when it comes to this device. We're making a router, and because the CPU has already been settled on, there's no way for it to handle 4 SFP+ ports. And even if we chose one that supports what you want, then the device would get more way more expensive. And you see where the sweet spot is.
@@tomazzaman can i do proxmox on it
i fan live with 2
i want to do HA for it this is why
Please dont sell it as a bare PCB !
Any particular reason? If it were sold both with and without an enclosure?
@@tomazzaman if it is both then there is no problem. But there is a very high probability of people using it without the enclosure leading to early deaths of the device due to dust/cat attacks etc.
only the pcb would be nice and everyone can print the case themselfs :) idea
Do you have a website?
and a discord server?
Not yet - will announce once either is ready. I'm currently quite overwhelmed with the planning stage! :)
Hey tomaž if your looking for ppl and what do do with their gear i would be willing to share my story 😅
I just looked at the intel n100 and its at 55$. Sure it has only 2 gigabit ethernet but it has 9 pcie 3.0 lanes.
N100 has no embedded options available. Meaning Intel doesn't guarantee any amount of time of production. So it might get deprecated a few years from now, going against what this router is all about (longevity).
Plus, as you mention, doesn't have any 10Gb networking interfaces, so we'd need to add those chips, likely getting close or over what the NXP CPU costs.
And finally, likely most importantly, NXP CPU is highly optimized for networking. Meaning it'll be faster and way more efficient at networking tasks than a general purpose N100.
Make the router faster with 4x2.5g, usb port to attach hdd for simple nas. more feature, more capabilities. 2 Copper 10g ports, wifi.
CWWK and their ilk (Topton, Minisforum, Beelink, etc) have the "do-everything miniPC" niche pretty well sewn up. A genuinely 10gig-capable router+firewall more or less doesn't exist sub-enterprise-tier, so he's got a lot better opportunity there imo. Hell, *I* couldn't justify it currently, but even at $750 he'd be firmly the cheapest thing in his performance/capability segment- and if he can swing a solid VPP implementation to open up future upgrade tiers with 25 or even 40gbps nics without needing dedicated ASICS or massively overpowered general-purpose CPUs, well, the sky's the limit.
Granted, those are a wide selection of medium-to-large-sized "ifs". But I'm real interested to see how it turns out for him.
docsis 4.0 arriving soon. The router needs to beat all specs. @@abroughear
Okay paused. So I would try to make different variants of the router, cheaper and premium.
Anything more than $380 and it'd have to offer something compelling that my UDM Pro does not have.
How about full access to the hardware, own choice of operating system, tons of documentation? :)
From having been in startups and a few product stints, beware of survivorship bias in your surveys. The problem with relying purely on followers and subscribers is their inbound marketing nature. You won't get feedback from the demographic that might actually be the true power users and even influencers of their communities, leaving you without feedback that you need to steer your project correctly.
Yup, agreed! I've done a lot of 1-on-1 research with some people from my personal network and the numbers were significantly higher. Last phone call I had, the other person literally said: "Price it 600, and I'll take two!" :)
@@tomazzaman+1 to that. Take a look at some stats for current high end routers, especially on enterprise products. Back when I still ran game and email servers, the challenge of keeping a router up was ventilation. If I wanted 24/7 cooling for a router with SoC churning at 15W, I had to get a server 1U product at least, which costs significantly more. It's worth spending some time to look through current offerings from various places like Fortinet, Palo Alto, Dell, etc.
Meanwhile, Apple took the answers for their products and tripled them... and people still buy it.
that is because of percived value of brand (i dont get it but OK)
You really should focus on the router part and the advance home user market. Personally I will have a separate switch anyway (I will need more than 3-4 ports), but what I was looking for is just a reasonably priced 10GB router with at least 2 10GB Base-T RJ45 Ports (WAN/LAN) because most Fiber-Modems have a RJ45-Out and there are already decent priced RJ45-10GB-Switches. But there are no “cheap” 10 GB routers for the home market. And RJ45 makes it all easier for the (advanced) home user. Why SFP+ ports? You shouldn’t try to compete with professional products. If SFP+ ports, at least make 2 & 2 RJ45/SFP+ combo ports (2 WAN/2 LAN)
I hear you loud and clear, but I don't agree with the last bit - I shouldn't compete with professional products? Why not? Because I definitely will. This router will be as professional as they come.
Honestly, I don’t think you will have any chance to enter the professional market. No offense. But this is just another game. But bringing value to the advanced homelab users with 10GB Fiber connection could be your niche.
10G for the advanced homelab users is my niche. But that doesn't mean the product won't be professional. No offense taken, I think it's the language barrier.
If you can bring a professional 10GB router to the homelab community for about +/- 500€ (or maybe even less) I am all in. But please with at least 2 10GB RJ45 ports ;-)
The survey limited us to $100 as the minimum. I wanted to say $50 but it wouldn't accept an answer below $100
I would employ another youtuber they watch to give the product a favorable review. Regarding the cost, I'd find and butcher some existing cheap piece that is close enough to use as a starting point. If it's not broken, don't fix it.
Nisem sledil čisto temu projektu ampak, po mojem Wifi-ja se ne rabi. Samo ustreliš se v nogo, ker ljudje potem pričakujejo nek hud performance, ki ga ni in je slaba volja.
Kaj pa kakšn extruded aluminium?
IMO je storage lahko 32GB
Storage bi lahko bil samo M.2 slot
Nabavne cene moraš po mojem še dodatno dol zbit, razmišlat o kakšnem Kickstarterju in si zadat višji cilj kot 1k enot.
Ram je predrag. A 4GB ne bi bilo dovolj?
Kaj se pričakuje, da bo uporabnik naložil na to zadevo? Nek OPNsense, PFsense? Po mojem je 4GB dovolj. Naj bo več rama za verzijo 2.0 al pa Pro verzijo :)
Sicer pa kul projekt. Upam, da ti uspe :)
PS:
Osebno sploh ne bi SFPjev dajal, ker 10Gbps je mal huda za to zadevo. Za neke dolge razdalje tud najbrž ni namenjen, da bi optiko fural. Za se izognit ISPjevi opremi pa rajši switch uporabiš, ker imaš najbrž še multicast, ki ga nočeš čez to zadevo pošiljat. Oziroma, če že mora bit, bi dal kot optional addon, ki si ga lahko sam vgradi uporabnik. Tudi marža je višja na taki zadevi potem. Pa lahko pol dodaš še kakšen M.2 extension al pa podobno, lahko tud kakšen SATA controller, če bi kdo hotu naredit potem kakšne vrste NAS iz tega.
LP, Tomaž (Tudi tukaj) ;)
Wifi-ja v originalno ne bo, bo pa mini-PCIe port, da se ga lahko doda kasneje. Extruded alu je grd, vsaj za moj okus. Sheet metal bo. Kar se pa portov tice, pa je precej koncen spec (2x SFP+ in 3x 2.5G RJ-45), ker smo z izbiro procesorja zakljucili. Gor bo najprej OpenWRT, ker je najbolj podprt, kasneje upam, da uspe nalozit OPNsense.
Apple proved that this method is no longer valid, you can continually lower the quality, limit what the customer can do with the device while increasing the price!
100 dolars was the minimum.
Questions answered by random people:
1: -, i get a router for free from my ISP
2: 1-2 months of my internet bill
3: 4-6 months of my internet bill
I know personally I would till want a 10g copper connection but I’m also not going buy this so my input doesn’t matter
I appreciate the feedback nonetheless. And the good part about SFP+ cages is that you can actually put in an RJ-45 transciever, turning it into a copper connection! 💪
@@tomazzaman yea but it is an added cost but that makes sense. I hope this goes Weill for u. I found your channel from the isp router video that just blew up
Sponsored by Calvin Klein?
Don't tell anyone!
Xiaomi is expanding its lineup of WiFi 7 routers with the new BE 3600, set to launch on January 30th in China. While the official price remains under wraps, pre-reservations are now open which shows the price will not be more than 299 yuan (around $43). very professional to make something that costs 10x more . applause!
In a race to the bottom, everyone loses.
😂 yeah I would like to get it for free. Let’s talk about that topic.
The problem, I think, with this method is that people have different incomes/budgets and "expensive" is relative/subjective heh
And people wanting to pay less than the hardware actually costs is a problem with capitalism heh
This is exactly why I will never become an inventor heh
That's the beauty of data. The more points, the clearer the picture. And less subjectivity.
@@tomazzaman "Numbers never lie but people do"
With a nearly 500$ price, and removing the original case, you are now in the price point of a dreammachine special edition from ubiquiti. Wich have integration with software and harware ecosystem, and have a descent support and waranty. I like your project, i will enjoy watching your video, but I will not buy this product.
Lets be honest anybody willing to pay less than 600 has no idea how expensive any 10Gbit purpose built firewall is.
I mean look at netgate or opensense hardware, bare minumum for a 1-2 port 10Gbit FW is 800+
And stuff from "proper" brands is at least 1k, usually significantly more than 1k.
And if you want to build your own FW, and don't want to go with 5+ year old, loud/ less efficient hardware you'll have still have it hard to stay below 600.
Yep, I don’t have much to add. I’m doing my best to optimize every single bit to bring the costs down but still, 10Gb requires some muscle, and muscle doesn’t come cheap.
Here is what i already wrote on previous iteration. Im one of those who said way to little but that is mainly because of im without a job(used to work as a programmer) so i value everything as gorenc ;).
you will know my idiotic anwser by my email as i have custom domain (same as my youtube username) when i remove my current situation out of the equasion i can say that i would be willing to pay 500+€ for it but less than mac mini. if you want to ask questions im available(clearly since i do not have a job) and you already have my email :) so just send the questions my way
PS. as a fellow youtuber im rewatching your video for algorithms sake
Appreciate both the feedback and the re-watch! And maybe, just maybe, gorenc can get a discount :D
@@tomazzaman well if my wish comes true ill have a job soon (been almost a year now), but gorenc in me apriciates the thought :)