Chapter 4: Advanced WAN Services Design Considerations (Part04)

SLA Overview

The SLA is a statement of intent from the provider.

The SLA should set customer expectations by specifying what service levels the service provider will deliver to the customer.

A good SLA covers common managed-service metrics:

• Mean time to repair (MTTR): MTTR is a measure of how long it takes to repair failures.

• Mean time between failures (MTBF): MTBF is a measure of how often failures occur. MTBF can be used to project how often failures are expected

Outage severity and causes vary. For many customers, it is useful to keep a log that tracks major, intermediate, and minor outages; the date and time of the outage; the time the service was restored; and the root cause of the outage. By tracking outages and noting the clarity of communication and the causes of outages, customers have concrete information to discuss at contract renewal time.

A good SLA also includes technical metrics:

• Packet loss: Packet loss occurs when one or more packets traveling across a network are lost and fail to reach their destination.

• Latency: Latency or delay is a measure of how much time it takes for a packet to get from one point in the network to another. Delay can be measured in either one-way or round-trip delay.

• Jitter: Jitter is the variable delay that can occur between successive packets. Based on network congestion, improper queuing, or configuration errors, the delay between each packet in a data stream can vary instead of remaining constant.

• IP availability: IP availability is a measure of the availability of IP services end-to-end across a network. TCP session breakage occurs after a component failure when the combination of convergence at Layer 2, Layer 3, and Layer 4 components causes a TCP session to fail.

Traffic throughput and general usefulness of a WAN link depend on very low packet loss rates (for example, 0.001 percent or lower), reasonably low latency, and low jitter. What these numbers should be depends on the type of traffic. VoIP requires good values for all. Best-effort Internet traffic is typically much more tolerant of latency and jitter. Layer 4 components such as Windows Server and Windows XP client stacks have about a nine-second tolerance for interruptions in IP availability.

Customers should avoid SLAs that average results across many devices or links. When averaging availability across enough routers, a device can be down for days and not violate some availability SLAs. Generally, customers want an SLA that covers the extreme behavior of any individual link or device.

SLA Monitoring

Customers should monitor the network status to track how well the service provider is doing at meeting the terms of the SLA.

Networks measurements let customers know what kind of service they are actually receiving. Measurements also provide a baseline for comparison when there are application issues. Measurements can help diagnose outages and performance problems, especially when it is unclear whether the problem is internal to the customer or with the service providers. Customers who receive and monitor alarms can help document the delay in notification or response from the provider after outages.

Some service providers allow their customers to view SLA performance data through a web portal. These measurements should be monitored as a way to track issues such as slow service due to congestion and to understand the health of the network. Even when service provider reports are accessible, many organizations also implement internal monitoring of key links. Internal monitoring can serve to activate an early response to a major outage, and can provide a cross-check that the service provider reports are not censoring the data shown.

Problems can arise if customers think they have been getting poor service because of violations of the SLA. Internal customer measurements can provide evidence of network issues to discuss with the service provider, especially if the service provider has not actively been measuring the service.

Summary

Underlying optical technologies such as SONET/SDH, CWDM, DWDM, and RPR enable the service provider to provide advanced WAN services such as storage, content switching, web hosting, instant messaging, and security built on an Ethernet framework.

Service providers use Metro Ethernet and MPLS VPNs as a flexible transport architecture to offer high-speed WAN Ethernet services such as transparent LAN services to support enterprise customers.

WAN service design involves a partnership between the service provider and customer, where the service provider manages services for the customer on an ongoing basis.

References

For additional information, refer to these resources:

• Cisco Systems, Inc. “Cisco MDS 9000 Family Fabric Manager Configuration Guide, Release 3.x,” at http://www.cisco.com/en/US/products/ps5989/products_configuration_guide_book09186a0080667aa0.html

• Cisco Systems, Inc. “Cisco CDWM Solution at a Glance,” at http://www.cisco.com/application/pdf/en/us/guest/products/ps6575/c1031/cdccont_0900aecd803a53ea.pdf

• Cisco Systems, Inc. “Cisco CDWM GBIC and SFP Solution Data Sheet,” at http://www.cisco.com/application/pdf/en/us/guest/products/ps6575/c1650/ccmigration_09186a00801a557c.pdf

• Cisco Systems, Inc. Cisco ONS 15454 DWDM Engineering and Planning Guide, Release 7.x, Chapter 1, “DWDM Overview” at http://www.cisco.com/en/US/products/hw/optical/ps2006/products_technical_reference_chapter09186a0080651e3b.html

• Cisco Systems, Inc. Fundamentals of DWDM Technology, at http://www.cisco.com/univercd/cc/td/doc/product/mels/cm1500/dwdm/dwdm_ovr.htm

• Cisco Systems, Inc. “Virtual Private LAN Services (VPLS) Introduction,” at http://www.cisco.com/en/US/products/ps6648/products_ios_protocol_option_home.html

• Cisco Systems, Inc. “Resilient Packet Ring Feature Guide,” at http://www.cisco.com/en/US/products/sw/iosswrel/ps1829/products_feature_guide09186a008022261a.html

• IEEE 802 LAN/MAN Standards Committee. “IEEE 802.17 Resilient Packet Ring Working Group” at http://www.ieee802.org/17/

• The Internet Engineering Task Force. RFC 4761: Virtual Private LAN Service (VPLS) Using BGP for Auto-Discovery and Signaling, at http://www.ietf.org/rfc/rfc4761.txt

• The Internet Engineering Task Force. RFC 4762: Virtual Private LAN Service (VPLS) Using Label Distribution Protocol (LDP) Signaling, at http://www.ietf.org/rfc/rfc4762.txt