scope:

SCOPE AND PURPOSE

Availability is a key measure in Service Level Agreements (SLA) between service providers and their customers as well as their vendors and suppliers. Several availability efforts have been addressed by various standards bodies. They can be briefly described as follows.

In the area of network availability, the access availability of routers in IP-based networks has been defined in terms of the availability of customer-facing ports [T1.TR.78-2003] and, more recently, in terms of the availability of customer-facing line cards [ATIS-0100025].

In the area of service availability:

• The availability of Voice over IP (VoIP) services was defined in terms of a Defects Per Million (DPM) metric [A-0100008]. An estimation methodology outline is also provided in this standard.

• In the ITU-T, general definitions of IP service availability [Y.1540] and MPLS service availability [Y.1561] have been defined. In addition, Ethernet service availability definition work is also ongoing.

From the perspective of service availability, a service provider needs to determine how best to deliver the desired availability to customers based on requirements specified in SLAs. Customers who agree to SLA-specified availability requirements include:

• Government Agencies: Agencies at the Federal, state, and local governments typically buy several types of telecommunications services from service providers and hence, they specify many requirements including service availability. The US Federal Government also specifies general requirements that dictate how large service providers are supposed to provide basic telephone service to the population.

• Enterprise Customers: Large corporations, financial institutions, universities, etc, are examples of Enterprise Customers. Such customers agree to very specific requirements for the type of service desired. Service availability is a key requirement for such customers.

The delivery of specific services to such customers typically is classified as "Enterprise Services". A service provider can manage availability as follows:

• Design appropriate network resources with appropriate reliability features to deliver the required level of availability.

• Implement necessary measurement capabilities to monitor and estimate the availability of the customer's service over specified time periods (e.g., monthly, annually, etc).

Both actions require significant investment in terms of planning and cost. Service providers may choose to strongly address both options. Alternately, a service provider may choose to invest heavily in the planning stage whereby significant efforts are dedicated to ensure that sufficient resource allocation, coupled with appropriate level of diversity and redundancy, is built into the network design [ATTOFC] such that the desired service availability can be satisfactorily achieved. This may mitigate the need for sophisticated and expensive monitoring and measurement capabilities for estimating service availability on the "back end".

This Technical Report provides an overview of the network resiliency design process for Enterprise Customers.