Web Services SLA: How to Optimize Service Level Agreement Performance

As IT moves beyond the stage of Web service availability and deployment with respect to a given service, the organization needs to be able to assure that the Web services SLA (service-level agreement) for that service is being met. IT faces other challenges as well as the deployed service becomes available multiple consumer types. Different requirements apply to services as they begin to be consumed by multiple applications and users.

Different consumers will have specific requirements surrounding their own applications -- for example: the level of response time they will tolerate. And while this sort of requirement may be simple to express, one of the greatest challenges IT must overcome is the reality that any one consumer or user of a service has the potential to (dramatically) impact the others.

In an extreme example, a malicious consumer could initiate a denial-of-service attack. This consumer could actually make so many requests to the service that the service ends up going down -- preventing legitimate consumers from accessing it. While this is an extreme example, it has been know to occur.

More typical cases, however, involve a single consumer placing a heavy (peak) load on a service, demanding a lot of information it; this situation can obviously cause availability issues for the other consumers. The first step to resolving these types of issues is to be able to understand use -- consumer by consumer -- of the services on the network so that the organization can then prioritize how to use and allocate resources effectively. SOA infrastructure software often assists in this process, as well.

Ensuring the Web Services SLA is met by employing SOA infrastructure software

Optimizing Services to Improve Web Services SLA Performance

In addition, in order to ensure that the Web services SLA for a service is accomplished, IT needs a way to understand how that service is actually being used. This capability is necessary in order to optimize the service, resolve issues and to detect anomalies in the first place. For example, imagine a credit-score lookup service. If there happens to be a high number of requests for credit scores emanating from a particular set of consumers, that situation may indicate that (potentially malicious) parties are trying to use the service in unintended or unauthorized ways.

Normally, this kind of information about consumers and their use of services can assist IT to understand such things as:

• When the services network is approaching the limits of its capacity
• When additional capacity will be required
• Which consumers are doing what sorts of operations when, and
• Which consumers, if any, are causing problems for others

Going further, IT can begin to use the discovered information about services and how they're used to understand and detect patterns of use.

SOA Management Provides Control, Backs Up Web Services SLA

In summary, as IT makes services available on a broad scale, the organization needs to understand that simply deploying services -- the process of Web service availability -- is not enough. IT needs the tools to ensure that services can be made available (safely) to diverse sets of consumers -- not only the anticipated, "planned-for" audience. IT must also be able to control exactly which information gets to each set of consumers. A SOA management infrastructure solution will provide this kind of fine-grained control. In addition, it can keep malicious consumers from improperly accessing and "breaking" information and functionality associated with services.

Find Out More

Trouble meeting your Web services SLA? Discover the secret to delivering on service-level agreements: download the free webinar, How Enterprises Can Leverage SOA to Share Information Securely