pressmaster - Fotolia

Beyond metrics, network SLAs should measure business ops

Enterprises need to ensure network service-level agreements provide a comprehensive view of network service performance to support business operations.

After a network outage or some other infraction, no organization is eager to receive service credits that service providers offer in their service-level agreement. Rather, enterprises look to their network SLA to guarantee acceptable levels of service when they sign a contract with their provider. The SLA also acts as an ongoing measure to ensure network services are performing sufficiently to meet their operational objectives.

Well-defined SLAs are an important piece of insurance for solid network performance. Unfortunately, SLAs are not always written in a way that reflects an end-to-end view of service quality. Too often, network SLAs concentrate on isolated elements of service delivery -- such as packet loss, latency, jitter and availability -- instead of focusing on the full picture.

As individual elements, these discrete SLA metrics are important to overall performance, but they don't give a complete view of network service quality. These isolated measures also don't contextualize how the service performance effectively supports business operations and the end-user experience.

What is a network SLA?

A network service-level agreement is a contract between a service provider and customer that outlines the expectation for infrastructure service quality. The contract delineates how network performance is measured. The network SLA also states how penalties and service credits are handled if performance levels are not met.

Network SLAs are built around metrics such as uptime, response time to an outage and mean time to repair (MTTR). Typically, service providers promise to deliver network availability of about 99.999%. This five-nines availability is the percentage of time a network component or service is accessible to a user in a given period, usually defined as a year.

If providers don't deliver that number -- which equates to about five minutes of downtime annually -- they offer service credits that cut the client's bill by a specified amount. Enterprises negotiate on how the value of these services decreased based on the severity of an outage or other performance issue.

Why is it important to have a network SLA?

At their most basic function, network SLAs ensure the service provider meets the network performance objectives of the client. Setting penalties for noncompliance with SLA metrics motivates the service provider to meet delivery expectations. SLAs also provide an important window for enterprises into the responsiveness of their service provider to any performance issues.

As operators have extended their portfolios to include more complex and strategic products in areas like cloud services and managed security, they reassessed their approaches to SLAs. Customers who want more accurate measures of service performance are pressing providers to offer SLAs that demonstrate service quality is sufficient to support business processes that run on the network.

Reacting to pressure from enterprise customers for SLAs that are better aligned to corporate objectives, service providers are retooling these agreements to focus less on component-level metrics to capture statistics that reflect the totality of network performance. These SLAs are created by measuring the performance of several service components -- often compiled from multiple services that support specific operations, such as an online credit card transaction processing system.

The objective of network SLAs is to build a model where the provider can accurately demonstrate how the managed services it delivers are supporting the business outcomes the client desires. The provider has to measure each component of each service. And the operator must correlate how each element contributes to or detracts from network service quality.

Further complicating this, some of the services associated with a particular business operation or process may be delivered by a partner. This means the provider may not have visibility into all the elements of that service.

What should the network SLA include?

A network SLA should define which services are delivered and the specifics about the following items:

  • Downtime.
  • IT service levels.
  • Response time to an outage.
  • MTTR.
  • Other components of network operations.

The network SLA should also specify the metrics against which the service quality is measured. It should also outline what service credits are offered if network performance falls short of the agreed-upon metrics.

The agreement should also identify what the service provider needs to deliver and support and which resources the client needs to maintain. For example, if a cloud service is involved, the service provider might be responsible for the connectivity component but not the assets in the cloud environment.

Several providers that deliver connectivity-related services offer network SLAs to enterprise customers, including cloud providers, telecom operators and systems integrators. These service providers also offer SLAs to one another when they supply network services to each other. For example, if AT&T is supplying connectivity services to AWS or Microsoft, it defines metrics in an SLA.

While providers haven't completely achieved the goal of putting the network service performance into business context, they are delivering better end-to-end SLAs. These give business customers a more accurate picture of true service performance as end users perceive it. SLAs also safeguard the interest of the service provider, level-setting delivery expectations and defining responsibilities.

Dig Deeper on Network infrastructure

Unified Communications
Mobile Computing
Data Center