High Availability Features for EX Series Switches Overview

 

High availability refers to the hardware and software components that provide redundancy and reliability for network communications. This topic covers the following high availability features of Juniper Networks EX Series Ethernet Switches:

VRRP

You can configure Virtual Router Redundancy Protocol (VRRP) for IP and IPv6 on most switch interfaces, including Gigabit Ethernet interfaces, high-speed Gigabit Ethernet uplink interfaces, and logical interfaces. When VRRP is configured, the switches act as virtual routing platforms. VRRP enables hosts on a LAN to make use of redundant routing platforms on that LAN without requiring more than the static configuration of a single default route on the hosts. The VRRP routing platforms share the IP address corresponding to the default route configured on the hosts. At any time, one of the VRRP routing platforms is the master (active) and the others are backups. If the master routing platform fails, one of the backup routing platforms becomes the new master, providing a virtual default routing platform and enabling traffic on the LAN to be routed without relying on a single routing platform. Using VRRP, a backup switch can take over a failed default switch within a few seconds. This is done with minimum loss of VRRP traffic and without any interaction with the hosts.

Graceful Protocol Restart

With standard implementations of routing protocols, any service interruption requires an affected switch to recalculate adjacencies with neighboring switches, restore routing table entries, and update other protocol-specific information. An unprotected restart of a switch can result in forwarding delays, route flapping, wait times stemming from protocol reconvergence, and even dropped packets. Graceful protocol restart enables a restarting switch and its neighbors to continue forwarding packets without disrupting network performance. Because neighboring switches assist in the restart (these neighbors are called helper switches), the restarting switch can quickly resume full operation without recalculating algorithms from scratch.

On the switches, graceful protocol restart can be applied to aggregate and static routes and for routing protocols (BGP, IS-IS, OSPF, and RIP).

Graceful protocol restart works similarly for the different routing protocols. The main benefits of graceful protocol restart are uninterrupted packet forwarding and temporary suppression of all routing protocol updates. Graceful protocol restart thus allows a switch to pass through intermediate convergence states that are hidden from the rest of the network. Most graceful restart implementations define two types of switches—the restarting switch and the helper switch. The restarting switch requires rapid restoration of forwarding state information so that it can resume the forwarding of network traffic. The helper switch assists the restarting switch in this process. Individual graceful restart configuration statements typically apply to either the restarting switch or the helper switch.

Redundant Routing Engines

Redundant Routing Engines are two Routing Engines that are installed in a switch or a Virtual Chassis. When a switch has two Routing Engines, one functions as the master, while the other stands by as a backup in case the master Routing Engine fails. When a Virtual Chassis has two Routing Engines, the switch in the master role functions as the master Routing Engine and the switch in the backup role functions as the backup Routing Engine. Redundant Routing Engines are supported on Juniper Networks EX6200 Ethernet Switches, Juniper Networks EX8200 Ethernet Switches, and on all EX Series Virtual Chassis configurations.

The master Routing Engine receives and transmits routing information, builds and maintains routing tables, communicates with interfaces and Packet Forwarding Engine components of the switch, and has full control over the control plane of the switch.

The backup Routing Engine stays in sync with the master Routing Engine in terms of protocol states, forwarding tables, and so forth. If the master becomes unavailable, the backup Routing Engine takes over the functions that the master Routing Engine performs.

Network reconvergence takes place more quickly on switches and on Virtual Chassis with redundant Routing Engines than on switches and on Virtual Chassis with a single Routing Engine.

Virtual Chassis

A Virtual Chassis is multiple switches connected together that operate as a single network entity. The advantages of connecting multiple switches into a Virtual Chassis include better-managed bandwidth at a network layer, simplified configuration and maintenance because multiple devices can be managed as a single device, a simplified Layer 2 network topology that minimizes or eliminates the need for loop prevention protocols such as Spanning Tree Protocol (STP), and improved fault tolerance and high availability. A Virtual Chassis improves high availability for the following reasons:

  • Dual Routing Engine support. A Virtual Chassis automatically has two Routing Engines—the switches in the master and backup routing-engine roles—and, therefore, provides more high availability options than standalone switches. Many high availability features, including graceful protocol restart, graceful Routing Engine switchover (GRES), nonstop software upgrade (NSSU), nonstop active routing (NSR), and nonstop bridging (NSB), are available for an EX Series Virtual Chassis that are not available on standalone EX Series switches.

  • Increased fault tolerance. You increase your fault tolerance options when you configure your EX Series switches into a Virtual Chassis. You can, for instance, configure interfaces into a link aggregation group (LAG) with member interfaces on different member switches in the same Virtual Chassis to ensure network traffic is received by a Virtual Chassis even when a switch or physical interface in the Virtual Chassis fails.

Juniper Networks EX2200 Ethernet Switches, Juniper Networks EX3300 Ethernet Switches, Juniper Networks EX4200 Ethernet Switches, Juniper Networks EX4300 Ethernet Switches, Juniper Networks EX4500 Ethernet Switches, Juniper Networks EX4550 Ethernet Switches, or Juniper Networks EX8200 Ethernet Switches can form a Virtual Chassis. EX4200, EX4500, and EX4550 switches can be interconnected together to form a mixed Virtual Chassis.

Graceful Routing Engine Switchover

You can configure graceful Routing Engine switchover (GRES) on a switch with redundant Routing Engines or on a Virtual Chassis, allowing control to switch from the master Routing Engine to the backup Routing Engine with minimal interruption to network communications. When you configure GRES, the backup Routing Engine automatically synchronizes with the master Routing Engine to preserve kernel state information and forwarding state. Any updates to the master Routing Engine are replicated to the backup Routing Engine as soon as they occur. If the kernel on the master Routing Engine stops operating, the master Routing Engine experiences a hardware failure, or the administrator initiates a manual switchover, mastership switches to the backup Routing Engine.

When the backup Routing Engine assumes mastership in a redundant failover configuration (that is, when GRES is not enabled), the Packet Forwarding Engines initialize their state to the boot state before they connect to the new master Routing Engine. In contrast, in a GRES configuration, the Packet Forwarding Engines do not reinitialize their state, but resynchronize their state to that of the new master Routing Engine. The interruption to traffic is minimal.

Link Aggregation

You can combine multiple physical Ethernet ports to form a logical point-to-point link, known as a link aggregation group (LAG) or bundle. A LAG provides more bandwidth than a single Ethernet link can provide. Additionally, link aggregation provides network redundancy by load-balancing traffic across all available links. If one of the links should fail, the system automatically load-balances traffic across all remaining links. In a Virtual Chassis, LAGs can be used to load-balance network traffic between member switches, which increases high availability by ensuring that network traffic is received by the Virtual Chassis even if a single interface fails for any reason.

The number of Ethernet interfaces you can include in a LAG and the number of LAGs you can configure on a switch depend on the switch model.

Nonstop Active Routing and Nonstop Bridging

Nonstop active routing (NSR) provides high availability in a switch with redundant Routing Engines by enabling transparent switchover of the Routing Engines without requiring restart of supported Layer 3 routing protocols. Both Routing Engines are fully active in processing protocol sessions, and so each can take over for the other. The switchover is transparent to neighbor routing devices, which do not detect that a change has occurred.

Nonstop bridging (NSB) provides the same mechanism for Layer 2 protocols. NSB provides high availability in a switch with redundant Routing Engines by enabling transparent switchover of the Routing Engines without requiring restart of supported Layer 2 protocols. Both Routing Engines are fully active in processing protocol sessions, and so each can take over for the other. The switchover is transparent to neighbor switching devices, which do not detect that a change has occurred.

To use NSR or NSB, you must also configure GRES.

Nonstop Software Upgrade

Nonstop software upgrade (NSSU) allows you to upgrade the software on a switch with dual Routing Engines or on a Virtual Chassis in an automated manner with minimal traffic disruption. NSSU takes advantage of GRES and NSR to enable upgrading the Junos OS version with no disruption to the control plane. In addition, NSSU minimizes traffic disruption by:

  • Upgrading line cards one at a time in an EX6200 switch, EX8200 switch, or EX8200 Virtual Chassis, permitting traffic to continue to flow through the line cards that are not being upgraded.

  • Upgrading member switches one at a time in all other Virtual Chassis, permitting traffic to continue to flow through the members that are not being upgraded.

By configuring LAGs such that the member links reside on different line cards or Virtual Chassis members, you can achieve minimal traffic disruption when performing an NSSU.

Redundant Power System

Most Juniper Networks Ethernet Switches have a built-in capability for redundant power supplies—therefore if one power supply fails on those switches, the other power supply takes over. However, EX2200 switches and EX3300 switches have only one internal fixed power supply. If an EX2200 switch or EX3300 switch is deployed in a critical situation, we recommend that you connect a Redundant Power System (RPS) to that switch to supply backup power if the internal power supply fails. RPS is not a primary power supply—it only provides backup power to switches when the single dedicated power supply fails. An RPS operates in parallel with the single dedicated power supplies of the switches connected to it and provides all connected switches enough power to support either Power over Ethernet (PoE) or non-PoE devices. For more information about RPS, see EX Series Redundant Power System Hardware Overview.