I've been getting lots of queries from my VC peers about our investment in Vyatta. For some reason or another, folks don't seem to want to believe that open source software and commodity hardware can go down the stack from the operating system instead of up the stack toward applications. It's a question that we asked ourselves when we first funded this company with our partners ComVentures and Arrowpath Venture Partners.
Are networking systems that different at their core than other complex applications running on commodity hardware these days?
In a few more words, the basic premise for the investment thesis was as follows: if open source can act as a disruptive technology at the operating system (Linux, SUSE, Debian, etc.), on complex applications such as databases (mySQL, postgres, etc.), CRM (SugarCRM) and middleware (JBoss) then why can't this same business paradigm translate to a network system called a router, firewall, VPN, IP PBX, load-balancer, etc.?
Sure, software applications don't have the same physical dependencies as networks in some situations, but when was the last time you added much to your network that wasn't T1, T3, OC3, or 10/100/1000 Ethernet? Who really installs IEEE802.5, SMDS, UltraNet, or even FDDI these days?
And, yes, commodity hardware with rotating media has a lower MTBF than a device running from NVRAM. That's why an appliance we're testing has the software running from RAM and loaded via compact Flash (the stuff that costs like $1/MB these days at BestBuy - commodity!). We took the rotating media out of the appliance entirely as a test and it's working fine.
Let's also remember the target market here. Clearly, it's not the core of ATT, UUNET, or anywhere else a CRS-1 or T-640 belongs. And, we're probably not going to be the best choice for your wiring closet where you put in a Cat6K with 192 ports of Gig Ethernet.
But, for everything in between - the branch office router, the enterprise firewall, the server load-balancer, the VPN concentrator, the IP PBX... well, there's a target market for an investment thesis. Besides, I'll bet that you have open source software on servers near to those devices anyway.