Apricot2015 No-Full Route Changed Our Lives @AS55394
?
0 likes?339 views
1) GREE was experiencing router crashes and long convergence times when faults occurred on their network due to running full BGP routing tables with 8 routers in a full mesh iBGP configuration. 2) The key factors identified were insufficient hardware capacity, increased route counts, and too many iBGP peers. 3) The best solution was to decrease the route counts by moving from full routes to partial routes plus default routes, reducing the routes from 400,000 to just 2,600. This solved the problems quickly without requiring new hardware or complex configuration changes.
Apricot2015 No-Full Route Changed Our Lives @AS55394
- 1. Copyright ? GREE, Inc. All Rights Reserved.Copyright ? GREE, Inc. All Rights Reserved. No-Full Route Changed Our Lives @AS55394 Osamu Kurokochi, ?Data Center Team, ?Infrastructure Headquarters
- 2. Copyright ? GREE, Inc. All Rights Reserved. Name ?Osamu Kurokochi Dept. ?Data Center Team, ? ?Infrastructure Headquarters, GREE, Inc? Self Introduction
- 3. Copyright ? GREE, Inc. All Rights Reserved. There ?were ?8 ?border ?gateway ?protocol ?(BGP) ?routers. Full ?route reception was in operation on all routers, ?and each BGP router was connected as iBGP ? peers in a full ?mesh topology. Summer ?in ?2012 ?(Configuration at the time) GREE environment R R R R R R R R
- 4. Copyright ? GREE, Inc. All Rights Reserved. Summer ?in ?2012 ?(occurrence of fault) One time, a fault on the Transit side occurred and caused peers ?to ?crash. Convergence of the routes at the time took time ???????and the Router ?CPU froze. iBGP ?peers also began to crash which caused chaos. A shutdown of about 5 minutes lasted intermittently until convergence. ?(The only thing we could do was to watch what happened.)
- 5. Copyright ? GREE, Inc. All Rights Reserved. Summer ?in ?2012 ?(cause of fault) There were 3 main factors. ?1. Insufficient hardware processing capability ?2. ?Increased number of routes ?3. ?Too ?many ?iBGP ?peers ?(not that many) Caused by one or a mixture of 3 ?factors above.
- 6. Copyright ? GREE, Inc. All Rights Reserved. Solution 1. ?Reinforced hardware ?Buy hardware having better performance. Solution 2. ?Configuration change ?Introduce ?RR and reduce the number of iBGP-\?? Peers. Solution 3. ?Decreased number of routes ?Decrease the number of routes with ?a ? mechanism ?to ?reduce ?the ?load ?during ?a ?BGP ?update. Breakthrough Solutions Considered at the Time
- 7. Copyright ? GREE, Inc. All Rights Reserved. Key Judgment Point Replacement of BGP ?routers at all bases also was considered but is difficult in terms of effort. When a ?procedure ?＾verification ?★ ?order ?★ ? ? ? ? ? ?delivery ★ ?maintenance arrangement￣ ? was considered, ?this remedy was too ?slow´ ? It was judged that the problem was difficult to solve by introducing new hardware.
- 8. Copyright ? GREE, Inc. All Rights Reserved. Key Judgment Point We narrowed the solutions down to ?solution 3. ?In our company?¨s business model, 99% ?of ? accesses ?were from mobile devices. ? ?The necessity of full ?route itself was ? reconsidered ?resulting ?in ?as ?follows: ?Full ?route ?★ ?Partial ?route ?+ ?Default ?route ?*Partial ?Route ?= ?3 ?domestic ?mobile carriers and 5 ?ASs.
- 9. Copyright ? GREE, Inc. All Rights Reserved. Transit Router Own router Transit Router 1 In-house filtering 2 TransitFilter method Full route Default ?route Partial ?route Default ?route RIB FIB Own router RIB FIB Partial ?Route Default ?Route Transparent GREE adopts this solution. There ?Are ?Two Partial Routes
- 10. Copyright ? GREE, Inc. All Rights Reserved. Summary of Solutions Solution ?1. ?Reinforced ?hardware ? Buy ?hardware ?having ?better ?performance. ? ?? ★ ?Veri?cation ?required ?and ?it ?takes ?time ?for ?delivery. Solution ?2. ?Con?guration ?change ? Implement ?RR ?to ?reduce ?the ?number ?of ?iBGP-\??Peers. ? ? ★ ?Veri?cation ?required, ?it ?takes ?time ?for ?delivery ?and ? no ?conclusive ?evidence ?that ?the ?problem ?will ?be ?recti?ed. Solution ?3. ?Decreased ?number ?of ?routes ? Lower ?the ?number ?of ?routes ?with ?a ?mechanism ?to ? reduce ?the ?load ?during ?a ?BGP ?update. ? ? ★ ?This ?can ?solve ?the ?problem ?in ?a ?short ?time ?and ?is ? reliable.
- 11. Copyright ? GREE, Inc. All Rights Reserved. Solution of Problem Number of routes: ?At the time, 400,000 routes ? ???★ ?reduced ?to ?approx. ?2600 routes Come on, line trouble! We actually tried the solution. These have been further reduced to approx. ?1800 ?routes.
- 12. Copyright ? GREE, Inc. All Rights Reserved. I ?thought... We have found that we can run ? ? ? ? ?our operations without full route. Why not reconsider whether you need it too?
- 13. Copyright ? GREE, Inc. All Rights Reserved.