BGP-4

BGP-4
Name	BGP-4
Developer	Internet Engineering Task Force
Initial release	1994
Latest release	RFC 4271
Type	Exterior gateway protocol
Transport	Transmission Control Protocol
Port	179

BGP-4 is a standardized exterior gateway routing protocol used to exchange routing and reachability information among autonomous systems on the Internet. It enables large-scale inter-domain routing by propagating route reachability, policy, and path information across networks operated by organizations such as AT&T, Verizon Communications, Amazon Web Services, Google LLC, and Cloudflare. BGP-4 is specified in technical documents produced by the Internet Engineering Task Force and implemented in routers from vendors including Cisco Systems, Juniper Networks, Arista Networks, and Huawei.

Overview

BGP-4 operates between autonomous systems identified by Autonomous System Numbers and supports policies used by network operators at entities like Level 3 Communications and NTT Communications. It is commonly run on routers from manufacturers such as Cisco Systems and Juniper Networks and is integral to the routing fabric connecting major exchanges like DE-CIX, LINX, AMS-IX, and cloud providers including Microsoft Azure and Google Cloud Platform. Administrators in organizations such as Facebook, Twitter, and Netflix use BGP-4 for traffic engineering, multihoming, and route control.

Protocol Operation

BGP-4 uses Transmission Control Protocol sessions over TCP port 179 to form peering between routers at companies such as AT&T, Comcast, and Verizon Communications. Sessions are established between edge routers at networks participating in ecosystems like Internet2, content delivery networks such as Akamai Technologies, and research networks like CESNET. BGP-4 exchanges routing information using messages that advertise prefixes associated with organizations such as IBM, Oracle Corporation, and Salesforce. Operators configure session parameters referencing operational guides from vendors like Cisco Systems and routing policies reflecting commercial arrangements with entities such as Level 3 Communications.

Path Selection and Attributes

Route selection in BGP-4 relies on a sequence of path attributes including AS_PATH, NEXT_HOP, LOCAL_PREF, and MULTI_EXIT_DISC as described in documents from the Internet Engineering Task Force. Decisions are influenced by peering relationships among providers such as Cogent Communications, CenturyLink, and Tata Communications as well as content networks including Cloudflare and Akamai Technologies. Techniques for traffic engineering reference vendor features from Juniper Networks, Arista Networks, and software projects like FRRouting and Quagga used by operators at DigitalOcean and Linode.

Messages and Finite State Machine

BGP-4 defines message types—OPEN, UPDATE, NOTIFICATION, and KEEPALIVE—used to manage sessions similar to protocols outlined in RFC 4271 maintained by the Internet Engineering Task Force. The finite state machine transitions—Idle, Connect, Active, OpenSent, OpenConfirm, Established—are part of operational training from vendors such as Cisco Systems and are fundamental to automated control logic in network controllers from Juniper Networks and orchestration platforms used by Amazon Web Services and Google Cloud Platform. Implementations in open-source projects like BIRD and FRRouting mirror behavior tested in carrier labs at NTT Communications and peering points such as LINX.

Implementation and Deployment

Major router vendors implement BGP-4 in platforms from Cisco Systems, Juniper Networks, Arista Networks, Huawei, and Hewlett Packard Enterprise. Service providers including AT&T, Verizon Communications, Sprint Corporation, and CenturyLink deploy BGP-4 for inter-domain routing, while cloud providers such as Amazon Web Services, Google Cloud Platform, and Microsoft Azure use it for hybrid connectivity and peering. Open-source routing stacks like BIRD, FRRouting, and Quagga are used by academic networks including Internet2 and research institutions like CERN and MIT.

Security and Extensions

Security concerns led to extensions and practices such as Resource Public Key Infrastructure (RPKI) developed by organizations like the Internet Society, route origin validation used by carriers including Level 3 Communications and Cogent Communications, and the BGPsec specification produced by the Internet Engineering Task Force. Operational mitigations include prefix filtering by operators at exchanges like DE-CIX, route registries maintained by registries such as RIPE NCC, ARIN, and APNIC, and monitoring services operated by vendors like ThousandEyes and Dyn (now part of Oracle Corporation). Research from universities such as Stanford University and University of California, Berkeley has influenced security tooling and anomaly detection used by networks including Facebook and Google LLC.

History and Development

BGP-4 evolved from earlier versions specified in documents by the Internet Engineering Task Force to support CIDR and Classless Inter-Domain Routing adopted in coordination with organizations like IANA and regional registries such as RIPE NCC and ARIN. Its development involved contributions from industry players including Cisco Systems, academic researchers at MIT, and operational input from backbone operators like Verizon Communications and AT&T. Subsequent updates and best current practices have been published as RFCs by the Internet Engineering Task Force and debated in community venues such as IETF meetings and at peering forums like Euro-IX.

Category:Internet protocols