[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-09-04。"],[],[],null,["# How Cloud Router works\n======================\n\nCloud Router is an API abstraction implemented by multiple and redundant BGP\ntasks, a dynamic route control plane, and Virtual Private Cloud (VPC) network\ncontrol and data planes. Understanding how these three software components work\ntogether helps you understand Cloud Router operations and how\nlearned-route-best-path-selection options work.\n\nSoftware components of Cloud Router\n-----------------------------------\n\nThere are several software components within Cloud Router and\nVPC:\n\nCloud Router BGP task\n: Cloud Router BGP tasks are grouped together within a region. Each\n BGP task communicates with a dynamic route control plane for its region and\n group. BGP tasks don't handle packet data processing. Instead, BGP tasks manage\n BGP sessions to send and receive BGP prefixes.\n\nDynamic route control plane\n: Each region contains a dynamic route control plane that communicates with\n BGP tasks for its region and group. In global dynamic routing mode, dynamic\n route control planes in one region also communicate with dynamic route control\n planes in other regions. Each dynamic route control plane sends messages to the\n VPC network control plane.\n\nVPC network control and data planes\n\n: Google Cloud uses the [Andromeda network virtualization stack (PDF\n download)](https://www.usenix.org/system/files/conference/nsdi18/nsdi18-dalton.pdf)\n as the distributed control and data plane for VPC networking, and\n includes the following components:\n\n VPC network control plane\n : Each region contains a VPC network control plane that\n receives information from the groups of dynamic route control planes in\n their own region. Each VPC network control plane programs\n dynamic routes in receiving VPC networks. VPC\n network control planes also enforce dynamic route quotas.\n\n VPC network data plane\n : Each region contains a VPC network data plane that\n evaluates and implements dynamic routes using information from the\n VPC network control plane. The VPC network\n data plane performs packet forwarding.\n\nCloud Router BGP tasks\n----------------------\n\nThe following table shows how many BGP tasks a Cloud Router uses for\ncommon scenarios:\n\nSoftware maintenance\n--------------------\n\nCloud Router maintenance events release new features and improve\nreliability. During maintenance, new BGP tasks take over as BGP speakers and\nresponders. Before maintenance, the last BGP task notifies its peer router in\none of the following ways:\n\n- If the peer router supports graceful restart, Cloud Router sends a\n graceful restart notification (a TCP `FIN` packet).\n\n- If the peer router *doesn't* support graceful restart, Cloud Router\n sends a BGP `CEASE` notification to the peer router to terminate the BGP\n session.\n\nCloud Router maintenance events aren't announced in advance because\nmaintenance events are automatic and not disruptive, as long as the peer\nrouter supports graceful restart. Maintenance events are designed to complete in\nless than 120 seconds---thus, Cloud Router uses a [120 second\ngraceful restart\ntimer](/network-connectivity/docs/router/how-to/managing-bgp-timers#graceful-restart-timer). For\ninformation about how to find completed maintenance events, see [Identify router\nmaintenance events](/network-connectivity/docs/router/how-to/identifying-router-events).\n\nIf the peer router supports graceful restart, the peer router logs a graceful\nrestart event during Cloud Router maintenance. In accordance with\n[Section 4.2 of RFC\n4724](https://datatracker.ietf.org/doc/html/rfc4724#section-4.2), the peer\nrouter must honor the 120 second Cloud Router graceful restart timer,\npreserving learned routes and continuing to advertise routes, in the event that:\n\n- Cloud Router stops sending BGP keepalive packets.\n\n- *Applicable only when BFD is configured* : Cloud Router stops\n sending BFD packets. Consequently, the peer router *must* honor the BFD\n control plane independent bit value of `0` because Cloud Router\n uses a control plane *dependent* BFD implementation. For more information,\n see [graceful restart and\n BFD](/network-connectivity/docs/router/concepts/bfd#graceful-restart-and-bfd).\n\nIf the peer router *doesn't* support graceful restart or if a peer router has\ngraceful restart disabled, Cloud Router sends a BGP `CEASE`\nnotification following [Section 4.5 of RFC\n4271](https://datatracker.ietf.org/doc/html/rfc4271#section-4.5). After the\n`CEASE` notification, the BGP session remains down until Cloud Router\nreplaces the BGP task. Making adjustments to the Cloud Router hold\ntimer or the peer router hold timer doesn't prevent the BGP session from\nterminating.\n\nPlanned Cloud Interconnect maintenance\n--------------------------------------\n\nFor [planned Cloud Interconnect\nmaintenance](/network-connectivity/docs/interconnect/support/infrastructure-maintenance-events#planned-maintenance),\nCloud Router sends a BGP `CEASE` notification that terminates the BGP\nsession, removing the session's learned and advertised routes. Neither the\ngraceful restart timer nor the negotiated BGP hold timer apply during planned\nmaintenance events.\n\nUnexpected BGP task failures\n----------------------------\n\nCloud Router uses multiple BGP tasks so that HA VPN\ntunnel pairs, Router appliances, and VLAN attachments that meet a\nCloud Interconnect SLA don't depend on a single BGP task. For more\ninformation, see the [Cloud Router BGP tasks](#bgp-tasks) section of\nthis document. If a Cloud Router BGP task fails unexpectedly,\nCloud Router isn't able to send one of the notifications that it\nnormally sends during software maintenance. However, both learned and advertised\nroutes remain for the duration of the negotiated [hold\ntimer](/network-connectivity/docs/router/how-to/managing-bgp-timers#hold_timer)."]]