His analysis of the control push size did not go to 1 billion. That is why I stated that for any constrained environment push would work. I do not know of a use case for 1 billion entities in those constraints.Internet-scale is a different problem. Yours,Joel
Sent via the Samsung Galaxy S® 6, an AT&T 4G LTE smartphone -------- Original message --------From: Dino Farinacci <farina...@gmail.com> Date: 3/16/18 20:44 (GMT-05:00) To: "Joel M. Halpern" <j...@joelhalpern.com> Cc: Tom Herbert <t...@quantonium.net>, David Meyer <d...@1-4-5.net>, "Alberto Rodriguez Natal (natal)" <na...@cisco.com>, Florin Coras <fcoras.li...@gmail.com>, i...@ietf.org, lisp@ietf.org, Eliot Lear <el...@cisco.com> Subject: Re: [Ila] [lisp] LISP for ILA - scaling Copying Eliot. I don’t remember Eliot analyzing the data-plane. But he did see how far the mapping database could scale with push and don’t recall he saying 1 billion would be achievable either. Dino > On Mar 16, 2018, at 2:53 PM, Joel M. Halpern <j...@joelhalpern.com> wrote: > > From the analysis Eliot did many years ago for a LISP push solution, for any > constrained solution (a data center, a mobile operator, a fixed service > operator) the number of entries is probably not a problem. Even for a > conventional router. Churn rate, in its various manifestations, could well > be an issue. > > Sharding is but one of several ways to divide and conquer to avoid those > issues. Separating control load from data plane load is also a useful way to > help keep things manageable. > > Yours, > Joel > > On 3/16/18 3:33 PM, Tom Herbert wrote: >> On Fri, Mar 16, 2018 at 12:17 PM, Dino Farinacci <farina...@gmail.com> wrote: >>> Yes, understand. But even in your constrained “domain”, there may be just >>> too much state to push to all nodes. Especially in the 5G use-case. It >>> wasn’t a problem in the LISP beta network because the proxy xTRs had >>> relatively coarse prefixes that reached lots of EIDs. >>> >> The state would need to be sharded. You'd probably need to do this >> anyway for mapping-servers or high thoughput Internet facing routers >> for which using a cache would be challenging. >> Tom >>>> require provisioning ILA-Rs to handle the full load if necessary to be >>>> robust. >>> >>> Yes indeed. >>> >>> Dino >>> >> _______________________________________________ >> ila mailing list >> i...@ietf.org >> https://www.ietf.org/mailman/listinfo/ila
_______________________________________________ lisp mailing list lisp@ietf.org https://www.ietf.org/mailman/listinfo/lisp