4 ms·

FreeGate? Wasn't that the company that thought the way to build a manageable router appliance was to take FreeBSD and rewrite all the software to use a custom b

FreeGate? Wasn't that the company that thought the way to build a manageable router appliance was to take FreeBSD and rewrite all the software to use a custom built database instead of their existing config files (without /etc/passwd, /etc/hosts, etc)? Didn't that mean replacing a large body of stable working code and engineering knowledge with fresh stuff with unknown behavior possibly understood by no more than one or two people in the entire world? Didn't that cause recruiting headaches, quality issues and schedule slips causing them to almost entirely miss the window for their product? I'd think it a bit disingenuous to call that company a success. If anything they should serve as an example of what happens when a company's technical leadership identifies the wrong problem to solve. Edit: there appears to be a timeline here http://www.freegate.net.au/news/press_releases.html http://www.freegate.net.au/news/press_releases.html with plenty of mention of funding and partnerships starting in 1997 but no announcements of actual shipments or sales.

ChuckMcM 14y ago
Throw away accounts with snarky names aside [1], the problem FreeGate set out to solve was that small to medium businesses wanted to be on the Internet, ISPs wanted to sell them Internet service, but the economics of doing that for non-tech savvy businesses was broken. This was in 1996. At that time, if you were Pizza hut, and your Internet access went down, and then you called support, they would likely as not ask you something like "can you ping the router on our end?" If the telephone company was run like that their question would be like "Can you measure the voltage between ring and tip?" We talked to a lot of ISPs, and a number of value added resellers (VARs), and small business (SMB) owners, and networking companies, the common bit that was underpinning a huge number of issues was that SMB's didn't have the resources to hire a system administrator or IT guy, the kinds of things they wanted to do however could mostly be automated. The ISPs wanted to sell to SMBs but if they called support too often it cut into their margins so much that they started losing money. The VARs were selling switches and patch panels but the 'big money' was above them selling servers and routers, they wanted to capture more of that business. The solution then was a system that could provide a full point of presence on the network and be fully debuggable by a service tech without having to roll a truck to the customer site or ask them questions they had no way of knowing how to answer. This has been a remarkably successful tool in the telephone market in the form of private branch exchanges (PBX). So how do you build a network server with the management and usability characteristics of a PBX? The path FreeGate chose was to design and build a 1U server that could fit inside the telephone racks of the time, and then create a management package on top of a stable OS release that would allow us to offer the local customer a UI for doing the kinds of things they wanted to do (like add or delete email accounts, put up web pages, or create a VPN tunnel between outlets or offices. FreeGate shipped the first one about 3 weeks later than the original schedule called for it to be shipped. That included designing a new motherboard, creating a chassis to hold it, and getting it through a bunch of modem qualification paperwork. As for market window, FreeGate, Whistle, and Cobalt who all ended up in variations of this space shipped about the same time. So I really don't think we'd characterize it as a market window 'miss.' We also didn't replace "large bodies of stable working code", we did create an entirely new management system, and we did create a way to proxy DNS requests so that you could have a DNS namespace for all of your machines both those with public addresses and those with private addresses. (the box did NAT and Firewalling as well). The biggest headache turned out to be Java. And more importantly how 'not true' the 'whole write once run everywhere mantra' was. Of course Microsoft and Netscape and Sun were all pointing fingers at everyone else but Java code that worked fine on Netscape didn't on IE and vice versa, and to make it worse from minor release to minor release of either of them. This comment: "Didn't that cause recruiting headaches, quality issues and schedule slips causing them to almost entirely miss the window for their product?" is amazingly exactly opposite reality. During the dot.com "boom" recruiting was challenging for everyone due to the insane competition for talent, people were giving away 1 year leases to a BMW sports car for sign on bonuses, kids with 1 year of experience out of school were demanding titles and pay of "architect." But none of that was at all due to our implementation. We did have a weird quality issue, it was too high. One of the strangest things I learned from that experience was that VARs (those people who re-sell gear from Cisco and Juniper etc) loved the fact that our product dropped in and worked, but they didn't like that it never broke. As it turned out their business model was predicated on making service calls and charging for each one. When they installed the FreeGate box the customer was 'done', they just didn't have issues. After the acquisition, the stock continued to gain value. I don't know about you, but being worth multiple millions of dollars (on paper of course) post acquisition made me feel pretty good about the exit. [1] Sidebar: This comment from 'freeflop' is showing as posted 8 hrs ago from an account created 8 hrs ago. This is not particularly unusual, especially when someone wants to talk about their own company in a bad light, but these things happened last century man. And it seems you still have a lot of pent up anger/hurt. I think if you came out in the open and had the discussion you might be able to get some closure but also respect your choice to live with it inside of you too.
- freeflop 14y ago
  From my perspective[1], instead of being a good example of a company that knew the right product to build and the right way to get it built, I think it just shows what happens when a company gets some bubble funding, tries to build "The Right Thing" and gets forced into an acquisition when it turns out too hard to do and the easy money runs out. I find your quality claims rather incredible but I have no experience with your box. However I do find it rather curious and ironic that you of all people, an original member of the Java team, were burned by quality problems in Java. I know WORA fooled a lot of newbies but shouldn't you of all people known enough about Java at that time to have avoided being bit by those? Perhaps you personally benefitted from the deal and I'm sure it was fun while it lasted, but your overall rationalization leaves me unsatisified. Let's be serious here. If FreeGate a. knew what to build b. knew how to build it we'd all know about it. [1] Well, I could have posted this from the HN account I abandoned 4 years ago but who cares. I'm just a FreeBSD contributor who got wind of what FreeGate was doing second hand and was a bit saddened when I heard about the whole database thing. That meant I thought they would be unlikely to ever be able to contribute back to the FreeBSD codebase in a meaningful way (e.g. the way Whistle did with netgraph).
  - ChuckMcM 14y ago
    FWIW we sold a few million units between the A1000 and the A500 over the lifetime of the product. The FreeBSD folks weren't interested in some "other" way to manage system configuration (it was all user level anyway), and third party awareness isn't a measure of quality for any product. To this comment "Well, I could have posted this from the HN account I abandoned 4 years ago but who cares." I don't know if anyone 'cares' whether you post on a newly created count or existing count, but I do care that creating a throwaway with a snarky name often can indicate a lot of unresolved personal pain and anger on the part of the poster. That stuff can eat away at you and leave you in a bad place. As a community we've lost too many good people to unresolved anger and depression.
    - freeflop 14y ago
      Well, people who misrepresent their part in successes and obscure or gloss over their role in failures bother me, especially when said individuals go on to offer advice such as yours about something as difficult as finding the right problem to solve and solving it in a practical way. I really don't have a dog in this particular race but anything that could substantiate those numbers or your other claims would be much more convincing and satisfying and go a long way to quelling my objections to your earlier comments.

ChuckMcM 14y ago

Throw away accounts with snarky names aside [1], the problem FreeGate set out to solve was that small to medium businesses wanted to be on the Internet, ISPs wanted to sell them Internet service, but the economics of doing that for non-tech savvy businesses was broken. This was in 1996. At that time, if you were Pizza hut, and your Internet access went down, and then you called support, they would likely as not ask you something like "can you ping the router on our end?" If the telephone company was run like that their question would be like "Can you measure the voltage between ring and tip?" We talked to a lot of ISPs, and a number of value added resellers (VARs), and small business (SMB) owners, and networking companies, the common bit that was underpinning a huge number of issues was that SMB's didn't have the resources to hire a system administrator or IT guy, the kinds of things they wanted to do however could mostly be automated. The ISPs wanted to sell to SMBs but if they called support too often it cut into their margins so much that they started losing money. The VARs were selling switches and patch panels but the 'big money' was above them selling servers and routers, they wanted to capture more of that business. The solution then was a system that could provide a full point of presence on the network and be fully debuggable by a service tech without having to roll a truck to the customer site or ask them questions they had no way of knowing how to answer. This has been a remarkably successful tool in the telephone market in the form of private branch exchanges (PBX). So how do you build a network server with the management and usability characteristics of a PBX? The path FreeGate chose was to design and build a 1U server that could fit inside the telephone racks of the time, and then create a management package on top of a stable OS release that would allow us to offer the local customer a UI for doing the kinds of things they wanted to do (like add or delete email accounts, put up web pages, or create a VPN tunnel between outlets or offices. FreeGate shipped the first one about 3 weeks later than the original schedule called for it to be shipped. That included designing a new motherboard, creating a chassis to hold it, and getting it through a bunch of modem qualification paperwork. As for market window, FreeGate, Whistle, and Cobalt who all ended up in variations of this space shipped about the same time. So I really don't think we'd characterize it as a market window 'miss.' We also didn't replace "large bodies of stable working code", we did create an entirely new management system, and we did create a way to proxy DNS requests so that you could have a DNS namespace for all of your machines both those with public addresses and those with private addresses. (the box did NAT and Firewalling as well). The biggest headache turned out to be Java. And more importantly how 'not true' the 'whole write once run everywhere mantra' was. Of course Microsoft and Netscape and Sun were all pointing fingers at everyone else but Java code that worked fine on Netscape didn't on IE and vice versa, and to make it worse from minor release to minor release of either of them. This comment: "Didn't that cause recruiting headaches, quality issues and schedule slips causing them to almost entirely miss the window for their product?" is amazingly exactly opposite reality. During the dot.com "boom" recruiting was challenging for everyone due to the insane competition for talent, people were giving away 1 year leases to a BMW sports car for sign on bonuses, kids with 1 year of experience out of school were demanding titles and pay of "architect." But none of that was at all due to our implementation. We did have a weird quality issue, it was too high. One of the strangest things I learned from that experience was that VARs (those people who re-sell gear from Cisco and Juniper etc) loved the fact that our product dropped in and worked, but they didn't like that it never broke. As it turned out their business model was predicated on making service calls and charging for each one. When they installed the FreeGate box the customer was 'done', they just didn't have issues. After the acquisition, the stock continued to gain value. I don't know about you, but being worth multiple millions of dollars (on paper of course) post acquisition made me feel pretty good about the exit. [1] Sidebar: This comment from 'freeflop' is showing as posted 8 hrs ago from an account created 8 hrs ago. This is not particularly unusual, especially when someone wants to talk about their own company in a bad light, but these things happened last century man. And it seems you still have a lot of pent up anger/hurt. I think if you came out in the open and had the discussion you might be able to get some closure but also respect your choice to live with it inside of you too.

freeflop 14y ago
From my perspective[1], instead of being a good example of a company that knew the right product to build and the right way to get it built, I think it just shows what happens when a company gets some bubble funding, tries to build "The Right Thing" and gets forced into an acquisition when it turns out too hard to do and the easy money runs out. I find your quality claims rather incredible but I have no experience with your box. However I do find it rather curious and ironic that you of all people, an original member of the Java team, were burned by quality problems in Java. I know WORA fooled a lot of newbies but shouldn't you of all people known enough about Java at that time to have avoided being bit by those? Perhaps you personally benefitted from the deal and I'm sure it was fun while it lasted, but your overall rationalization leaves me unsatisified. Let's be serious here. If FreeGate a. knew what to build b. knew how to build it we'd all know about it. [1] Well, I could have posted this from the HN account I abandoned 4 years ago but who cares. I'm just a FreeBSD contributor who got wind of what FreeGate was doing second hand and was a bit saddened when I heard about the whole database thing. That meant I thought they would be unlikely to ever be able to contribute back to the FreeBSD codebase in a meaningful way (e.g. the way Whistle did with netgraph).
- ChuckMcM 14y ago
  FWIW we sold a few million units between the A1000 and the A500 over the lifetime of the product. The FreeBSD folks weren't interested in some "other" way to manage system configuration (it was all user level anyway), and third party awareness isn't a measure of quality for any product. To this comment "Well, I could have posted this from the HN account I abandoned 4 years ago but who cares." I don't know if anyone 'cares' whether you post on a newly created count or existing count, but I do care that creating a throwaway with a snarky name often can indicate a lot of unresolved personal pain and anger on the part of the poster. That stuff can eat away at you and leave you in a bad place. As a community we've lost too many good people to unresolved anger and depression.
  - freeflop 14y ago
    Well, people who misrepresent their part in successes and obscure or gloss over their role in failures bother me, especially when said individuals go on to offer advice such as yours about something as difficult as finding the right problem to solve and solving it in a practical way. I really don't have a dog in this particular race but anything that could substantiate those numbers or your other claims would be much more convincing and satisfying and go a long way to quelling my objections to your earlier comments.