The ins and outs of service level management

Delivery of high-speed data services has become very competitive with customers demanding business-level service level agreements (SLAs) with the same quality of service (QoS) that is inherent in private line services.

The evolution of next-generation data services has also resulted in greater complexities for telecommunications service providers when addressing internal processes and functions such as ordering, provisioning, billing and customer care. When telecom service providers offered only traditional communication services, these functions were generally straightforward and could be managed in a stovepipe fashion. However, next-generation services such as DSL rely on the tight integration of multiple layered network technologies and use the inherent capabilities of each layer to provide customers with new services that can be offered with a guaranteed QoS.

Today's communications business customers require firm commitments in the form of legal contracts that cover all types of agreements for an ever-expanding range of services. They also demand service quality requirements with the expectation that they can "pick and choose" the options in the SLAs themselves.

To meet this challenge, many network service providers (NSPs) are providing extremely aggressive SLA terms. Unlike previous SLAs that focused on operational parameters such as availability, MTBF (mean time between failure), MTTR (mean time to repair) and service delivery time, NSPs must also now commit to service's performance including throughput, delay and loss parameters as part of the SLA contract (Figure 1).

In a recent analyst survey of NSP customers most said they would be willing to pay 30% more for guaranteed QoS than for best-effort service.

As a result, service quality management has become essential for service providers attempting to differentiate their network offerings. This paper describes the service level management process, what service level agreements cover and what service providers should look for when choosing among the many vendors offering solutions in this area.

An SLA is a contract between a service provider and a customer that specifies, in measurable terms, the levels of service that will be furnished. The metrics an SLA might measure can include the percentage of time the service will be available to the customer, the number of simultaneous users that can be served by the network, or specific performance benchmarks against which performance can be compared and monitored on a regular basis. In addition, the SLA contract must also stipulate what penalties may be applied if the service provider does not meet the SLA requirements or compensation if the customer uses more resources than originally agreed upon.

When a customer first contacts the service provider to request a new service or modify and existing service, they need to agree what Service Level parameters are to be included. SLA contract compliance should be monitored by measuring key performance indicators. A comprehensive set of guidelines have been defined by the TeleManagement Forum, resulting in the SLA handbook (GB917) and the Performance Reporting Concept and Definitions (TMF701) document, which categorize and detail measured key performance indicators including:

Network parameters

• Average/maximum/minimum bandwidth: Information units per second transferred

• Availability: The error-free available time as a percentage of the total connectivity time

• Latency: The propagation delay of a meaningful information unit

• Response Time: The time passing from a request for a service to a response to that request

• Jitter (latency variation): The variation of the delay between data units within the same service session or during a time frame

• Error Rate: The ratio of the rejected units of information to the total information the service attempts to deliver

• Sever Error Seconds: The amount of consecutive seconds within a session in which the error rate was higher than a certain threshold

• Accuracy: The ratio of correct content transferred by the services to the total contents

Operational parameters

• Mean time between failure

• Mean time to repair

• Mean time to provision the service

• Mean time to notify a customer about a service violation

Figure 2 depicts the relationship between KPIs and SLA as reflected in various service provider's current offerings.

A key element for effectively managing the service fulfillment processes is a comprehensive order management solution. An OM solution is ultimately responsible for:

1. Taking the customers service request order

2. Negotiating the corresponding SLA requirements and pricing plans

3. Verifying the availability and feasibility of the network to support the service

4. Managing the design, assign, and activation processes that need to take place to provision the service in the network, and finally

5. Notifying the billing system that the service is active.

Once the service has been successfully provisioned, compliance of the associated SLA contract then needs to be regularly monitored for any violations. When network failures or traffic congestion occur that result in the service provider failure to meet the guaranteed QoS, an alert should be initiated, a trouble ticket should be opened and, if necessary, an automated rebate/penalty should be assigned to the affected customer(s). The service provider then needs the necessary tools to prioritize the customers and restore the service as quickly as possible.

The OM solution must also be flexible enough to quickly and efficiently manage changes to the network, such as the introduction of new services and pricing plans, as well as ad-hoc changes to the SLA such as changing the QoS parameters.

The integrated process described above is known as service level management. Today's customers are demanding the highest standard/level of QoS, while on the other hand, expectations for lowered costs. With an increased number of competitive options, this paradox poses a complex obstacle for service providers to overcome.

One way organizations can drive down their costs is by transitioning from their private networks to the public network. In the banking industry, for example, most large banks use their own private networks for execution and monitoring of transactions, including fund withdrawals from automated teller machines and fund transfers from one account to the next. Private networks are used because the banks have confidence in the quality and security provided therein, which is something their customers, in turn, demand and expect.

However, developing and maintaining a private network is a costly proposition. Start-up costs alone require a significant fiscal commitment. Serving customers off the public network is a more cost-effective option. So how can a customer-facing company (i.e., a bank) be confident that the public network will adequately meet their requirements and the needs of its customers?

The use of virtual private networks, where multiple companies use a service provider's shared resource with guaranteed QoS and support for service level management, can help generate this level of confidence. The SLA contract is created to help guarantee the QoS, affording companies the opportunity to use the shared network and turn around cost savings to the customer.

The SLA also helps create safeguards for the customer, through definitions around responsibilities and requirements in the event of a lapse in service or failure on the service provider's part. This offers customers the security of knowing that potential problems will be identified and remedied quickly.

Ultimately, the goal of the a service level management is to guarantee that mission-critical applications be delivered to customers by way of the public network. Are we there yet? Most likely not. There are still too many quality issues pertaining to the public network that need to be addressed. However, as network technologies such as IP MPLS evolve to provide integrated support for traffic engineering and QoS, service providers can use them to support and offer SLAs. Customers will be able to move away from using private networks and use the shared service providers network instead. Key to this transition will be the availability of comprehensive service level management solutions that can effectively use and manage the QoS capabilities inherent in the network.

Moving now to the actual offerings of SLM solutions, service providers should demand the following from their SLM vendors:

• Flexibility and rapid definition of the contract template tailored for specific market segments

• A convergent solution which offers support for corporate customer hierarchies, complex products, contracts and pricing plans

• Intimate integration with the billing system and the CRM system

• Eligibility rules to ensure customer entitlement to the SLA and compatibility between the various services ordered

• One legal umbrella for all type of agreements

• Easy capturing and negotiation of complex contracts using a workflow engine and a friendly negotiation GUI

• Resolution of contract conflicts

• Notification prior to contact expiration

• One integrated solution instead of separate systems

• Support for multiple sales channels

• Proactive monitoring of contract compliance.

There are a number of small companies in today's market that specialize in SLA monitoring. However, few companies (i.e. Amdocs) offer SLM as part of an end-to-end integrated solution tailored exclusively for communications service providers, working alongside other key components and functions including CRM, billing, provisioning and order management.

In many ways, SLM is an outgrowth of CRM as it governs the quality of service the service provider is going to be able to provide the corporate customer and ultimately the end-user. From a marketing and sales perspective, SLAs can be sold as an added-value service and differentiator against a service provider's competition. From a customer service perspective, SLM allows the service provider to proactively notify key customers upon quality of service issues, thereby helping to avoid customer dissatisfaction. With effective integration with network-smart trouble ticketing, the QoS degradation can be fixed quickly and efficiently, minimizing the impact on the customer.

Furthermore, the current quality of service of each customer should be presented to the CSR as part of the integrated view of the customer. This will allow the CSR to fully understand the customer's experience during every interaction with the customer. Customers also expect SLA reporting to be available on the web, as part of their self-service channel.

For these and other reasons, it is essential for service providers to work with an SLM vendor who has a clear understanding of the communications industry and the complexities associated with its products and services.

It is also important for service providers to understand the relationship between the contract and the capturing of an order. The most efficient and timely system would be one where the order management functions (OM) were tied directly to the contract. This way, once it is decided which products and/or services are going to be ordered, the system would point each party directly to a contract already in place. Basically, when providing a product or service to its customers, the service provider automatically offers the SLA as well. Most, if not all, SLA specialty companies do not offer this type of automated direct tie-in to the order management function.

While it is essential to have the SLM solution governing the overall quality of service, it may be appropriate to have more specific SLAs designed for different components of the organizational structure. Vendors who have the ability to fully integrate themselves into the service providers internal workings - such as those who offer a full range of customer care and billing services - have the ability to provide a much more comprehensive SLA as well as monitor it more stringently and adjust it as required.

Doron Gover is Director of Business Development for Amdocs.

Visit Amdocs online.