- Table of Contents
- Related Documents
-
Title | Size | Download |
---|---|---|
01-High Availability Overview | 72.56 KB |
Communication interruptions can seriously affect widely-deployed value-added services such as IPTV and video conference. Therefore, the basic network infrastructures must be able to provide high availability.
There are three effective ways to improve availability:
· Increasing fault tolerance
· Speeding up fault recovery
· Reducing impact of faults on services
Availability requirements
Availability requirements fall into three levels based on purpose and implementation.
Table 1 Availability requirements
Level |
Purpose |
Implementation |
1 |
Decrease system software and hardware faults |
· Hardware—Simplifying circuit design, enhancing production techniques, and performing reliability tests. · Software—Reliability design and test. |
2 |
Protect system functions from being affected if faults occur |
Device and link redundancy and deployment of switchover strategies. |
3 |
Enable the system to recover as fast as possible |
Providing fault detection, diagnosis, isolation, and recovery technologies. |
The level 1 availability requirement should be considered during the design and production process of network devices. The level 2 availability requirement should be considered during network design. The level 3 availability requirement should be considered during network deployment according to the network infrastructure and service characteristics.
Availability evaluation
Typically, Mean Time Between Failures (MTBF) and Mean Time to Repair (MTTR) are used to evaluate the availability of a network.
MTBF
MTBF is the predicted elapsed time between inherent failures of a system during operation. It is typically expressed in hours. A higher MTBF means a higher availability.
MTTR
MTTR is the average time required to repair a failed system. MTTR in a broad sense also involves spare parts management and customer services.
MTTR = fault detection time + hardware replacement time + system initialization time + link recovery time + routing time + forwarding recovery time. A smaller value of each item, a smaller MTTR, and a higher availability.
High availability technologies
As previously mentioned, increasing MTBF or decreasing MTTR can enhance the availability of a network.
High availability technologies can be classified into fault detection technologies and protection switchover technologies.
Fault detection technologies
Table 2 Fault detection technologies
Technology |
Introduction |
Reference |
Ethernet OAM |
As a tool monitoring Layer 2 link status, Ethernet OAM is mainly used to address common link-related issues on the “last mile.” You can monitor the status of the point-to-point link between two directly connected devices by enabling Ethernet OAM on them. |
Chapter “Configuring Ethernet OAM” |
BFD |
Bidirectional forwarding detection (BFD) provides a single mechanism to quickly detect and monitor the connectivity of links or IP forwarding in networks. To improve network performance, devices must quickly detect communication failures to restore communication through backup paths as soon as possible. |
Chapter “Configuring BFD” |
NQA |
Network Quality Analyzer (NQA) analyzes network performance, services and service quality through sending test packets, and provides you with network performance and service quality parameters such as jitter, TCP connection delay, FTP connection delay and file transfer rate. |
Network Management and Monitoring Configuration Guide |
Monitor Link |
Monitor Link works together with Layer 2 topology protocols to adapt the up/down state of a downlink port to the state of an uplink port. This feature enables fast link switchover on a downstream device in response to the uplink state change on its upstream device. |
Chapter “Configuring Monitor Link” |
Track |
The track module is used to implement collaboration between different modules. The collaboration here involves three parts: the application modules, the track module, and the detection modules. These modules collaborate with one another through collaboration entries. That is, the detection modules trigger the application modules to perform certain operations through the track module. More specifically, the detection modules probe the link status, network performance and so on, and inform the application modules of the detection result through the track module. Upon aware of the changes of network status, the application modules deal with the changes accordingly to avoid communication interruption and network performance degradation. |
Chapter “Configuring track” |
Protection switchover technologies
Protection switchover technologies aim at recovering network faults. They back up hardware, link, routing, and service information for switchover in case of network failures to ensure continuity of network services.
Table 3 Protection switchover technologies
Technology |
Introduction |
Reference |
Active and Standby Switchover |
Devices supporting active and standby switchover are normally equipped with two main processing units (MPUs), with one being the active MPU, and the other being the standby MPU. The configurations on the standby MPU are the same as those on the active MPU. When the active MPU fails or is plugged out, the standby MPU automatically becomes the active MPU to ensure non-stop operation of the devices. |
Chapter “Configuring active and standby switchover” |
Link Aggregation |
Link aggregation aggregates multiple physical Ethernet links into one logical link to increase link bandwidth beyond the limits of any one single link. This logical link is called an aggregate link. It allows for link redundancy because the member physical links can dynamically back up one another. |
Layer 2—LAN Switching Configuration Guide |
Smart Link |
Smart Link is a feature developed to address the slow convergence issue with STP. It provides link redundancy as well as fast convergence in a dual uplink network, allowing the backup link to take over quickly when the primary link fails. |
Chapter “Configuring Smart Link” |
MSTP |
As a Layer 2 management protocol, the Multiple Spanning Tree Protocol (MSTP) eliminates Layer 2 loops by selectively blocking redundant links in a network, and in the mean time, allows for link redundancy. |
Layer 2—LAN Switching Configuration Guide |
RPR |
Resilient Packet Ring (RPR) is a new MAC layer protocol designed for transferring mass data services over MANs. It can operate on synchronous optical network/synchronous digital hierarchy (SONET/SDH), Dense Wavelength Division Multiplexing (DWDM) and Ethernet to provide flexible and efficient networking schemes for broadband IP MANs carriers. |
Chapter “Configuring RPR” |
FRR |
Fast Reroute (FRR) provides a quick per-link or per-node protection on an LSP. In this approach, once a link or node fails on a path, FRR comes up to reroute the path to a new link or node to bypass the failed link or node. This can happen as fast as 50 milliseconds thus minimizing data loss. Protocols such as RIP, OSPF, IS-IS, static routing, and RSVP-TE support this technology. |
Layer 3—IP Routing Configuration Guide, MPLS Configuration Guide/Configuration guide of the corresponding protocols |
GR |
Graceful Restart (GR) ensures the continuity of packet forwarding when a protocol, such as BFP, IS-IS, OSPF, LDP, or RSVP-TE, restarts or during an active/standby switchover process. It needs other devices to implement routing information backup and recovery. |
Layer 3—IP Routing Configuration Guide, MPLS Configuration Guide/Configuration guide of the corresponding protocols |
VRRP |
Virtual Router Redundancy Protocol (VRRP) is an error-tolerant protocol, which provides highly reliable default links on multicast and broadcast LANs such as Ethernet, avoiding network interruption due to failure of a single link. |
Chapter “Configuring VRRP” |
A single availability technology cannot solve all problems. Therefore, various availability technologies should be used to enhance network availability on a basis of detailed analysis of network environments and user requirements. In addition, network availability should be considered prior to building a network. For example, access-layer devices should be connected to distribution-layer devices over redundant links, and core-layer devices should be fully meshed.