On Wed, Oct 29, 2008 at 8:28 AM, Eliot Lear <[EMAIL PROTECTED]> wrote: >> On Wed, Oct 29, 2008 at 2:58 AM, Eliot Lear<[EMAIL PROTECTED]> wrote: >>> To put it another way, >>> home and SMB networks really don't even have an option today to be >>> multihomed (at least not at the network layer), >> >> Not correct. Anyone buying at least two T1s @ $600/mo or so each >> qualifies for a /24 for multihoming from one of the ISPs and an AS# >> from ARIN. > > They're going from $25 - $90 or so single homed (what most pay) to $1200 > multihomed plus a BGP capable router plus the cost of the ASN, and that's > only for failover. If you want to actually make use of optimal routing, how > much will that router cost? But all of that is NOTHING in comparison to the > costs of the expertise needed to manage such a service.
Hi Eliot, A Cisco 2811 with extra RAM capable of handling full BGP for two T1s costs about $2500. BGP-capable routers don't get expensive until you need one that can do both BGP and a few hundred mbps of traffic. As for expertise, the two most common configurations (primary/fail and balanced share) could be easily commoditized with the instructions: Here's your preconfigured router for the two ISPs you selected. If one ISP ever acts flakey, pull the T1 cord. The hobbyist without any expertise is going from the sub-$100 range to $1200. Or rather, for the most part he isn't because his Internet link isn't worth that much to him. The small business is going from $600 to $1200, which isn't so big a jump. > Who said anything about removing BGP? Certainly not me. Not sure what you > mean by "disrupting." My understanding is that APT replaces BGP entirely within its islands from the outset, Ivip replaces BGP following a transition period and LISP retasks BGP to carry routing for only the core ingress and egress nodes. Apologies to the authors if I have that wrong; it has been a long time since I read any of the three proposals and I could easily be crosslinking other discussions in my head. Can you point me to documentation that clarifies the situation? Anyway, a retasking of BGP would qualify as "disruptive" since BGP would cease to be usable for its original purpose of providing routing for enduser-level addresses. On Wed, Oct 29, 2008 at 10:06 AM, Scott Brim <[EMAIL PROTECTED]> wrote: > But the main point is whether small networks are managed differently > enough that they could be treated differently. Multihoming is one > factor, renumbering is another, and there are certainly others. Scott, I suspect that's mostly a red herring. It takes two to communicate. No matter how I manage my network, I only control half that process. When the agency I'm contracting for wants my source IP address to put in their firewall, I don't have the luxury of saying, "Gee, I really don't manage my network that way." Regards, Bill Herrin -- William D. Herrin ................ [EMAIL PROTECTED] [EMAIL PROTECTED] 3005 Crane Dr. ...................... Web: <http://bill.herrin.us/> Falls Church, VA 22042-3004 _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
