A system and method for end-to-end quality assurance between DCs based on SR and ESI

Abstract

The invention discloses a system for realizing end-to-end quality assurance between DCs based on SR and ESI, including several DC network ends, and three underlay networks are isolated between any two DC network ends; any one DC1 network end connects to another When the DC2 network end communicates the encapsulated VXLAN data packet, the data packet first enters the CE of the DC1 network end into the VTEP, then enters the DCI from the VTEP, and then enters the DCI of the DC2 network end through the underlay, and then enters the VTEP from the DCI. Finally, it enters the CE on the DC2 network side, and the communication is completed. In the underlay network between any two DCs, configure SR technology to realize fast rerouting and protect the access quality of the wide area network between DCs; in the overlay network, configure ESI technology in VTEP and DCI to realize DC to DC end End-to-end redundancy protection.

Description

technical field [0001] The invention relates to the field of network technology, in particular to a system and method for realizing end-to-end quality assurance between DCs based on SR and ESI. Background technique [0002] The VXLAN IETF Draft (vxlan standard draft) (draft-mahalingam-dutt-dcops-vxlan-01) stipulates that the communication between VTEPs (VXLAN Tunnel Endpoints, the endpoints of VXLAN tunnels) is based on the characteristics of the underlay network, that is, the deployed IGP / EBGP / Routing features of protocols such as MPLS. [0003] Routing features such as loop avoidance, fast rerouting, path selection, and load sharing for VXLAN data packets between VTEPs are based on the IGP / BGP features of the underlay network, which has the following two problems: [0004] The convergence speed of the IGP network is slow when the link is faulty. When the IGP network fails, even if fast detection tools such as BFD are deployed, the IGP network often needs more than 1000ms...

AI Technical Summary

Problems solved by technology

[0004] The convergence speed of the IGP network is slow when the link is faulty. When the IGP network fails, even if fast detection tools such as BFD are deployed, the IGP network often needs more than 1000ms to reach the second level during the path recalculation process, which cannot meet the requirements of the service without perception. ; When the underlying underlay network cannot achieve fast convergence or the VTEP node or DCI edge node in the DC fails, and the upper-layer overlay network lacks a fast and effective redundancy protection mechanism, rapid service switching cannot be achieved

In addition, the load balancing parameters of the VXLAN control plane are not defined in the VXLAN IETF Draft, so that the load balancing of CE between different VTEP nodes cannot be realized.

[0005] The traditional underlay network cannot select a suitable path according to the business attributes. It regards the VXLAN data packet as an ordinary IP packet. When forwarding between VTEPs, based on the HASH algorithm, it is randomly forwarded to a path. The delay, bandwidth, and jitter of the path and other parameters may not necessarily meet the business requirements of VXLAN data packets

Images