Ways to accelerate Web services

Web services environments suffer from architectural shortcomings that cause the technology’s performance to suffer in enterprise companies, says a recent report from Burton Group.

To start, Web services are based on a complex, distributed application infrastructure. As enterprise companies grow the number of Web services, the number of routing hops and processing nodes that Simple Object Access Process (SOAP) messages need to traverse to complete Web services transactions will also increase – which makes ensuring performance more difficult.

According to analyst James Kobielus, the most basic performance issues with SOAP-based Web services stem from constraints in the middleware, including content encodings, CPU-intensive application-server processes and unreliable messaging transports.

Yet SOAP-based application environments can take advantage of benefits such as heterogeneous interoperability, flexibility, adaptability, extensibility and reusability. Many more Web services performance problems can be attributed to suboptimal application designs, software configurations and hardware deployments than to shortcomings in Web services standards, Kobielus says.

But there is hope. Vendors provide a growing range of products and services that compress Web services traffic, offload CPU-intensive Web services functions from application servers, and reduce latencies of end-to-end Web services transactions. And enterprise companies are implementing new approaches to Web services scaling, acceleration, performance management and traffic management.

IT professionals can employ a broad range of approaches for managing Web services performance. Among them are connection and session monitoring; capacity expansion; co-processing; compression; caching; content-based routing; and client-side application logic deployment.

Connection- and session-monitoring technology will let IT managers measure the end-to-end performance of SOAP-based Web services sessions and transactions over network connections in real time, the report says. Capacity expansion involves adding Web services traffic processing power in strategic infrastructure locations. And co-processing offloads specialized functions from one or more Web or application servers to intermediary processors, which would prevent bottlenecks from occurring on overworked servers.

Compression reduce Web services traffic into smaller bit counts and send less data over network connections, speeding Web services delivery. Caching technology achieves similar results as compression but works a bit differently. Caching involves sending Web services content to one or more distribution points, from which client applications or browsers can retrieve and download the content.

Content-based routing involves executing a policy-driven flow of Web services traffic among two or more application components. Lastly, client-side application logic deployment works by running more functions on Web services clients, and – in the process – keeping unnecessary Web services workloads off the servers and unnecessary traffic off the network.

Kobielus adds that some of the scaling and acceleration technologies have proven successful already, but “no single Web services deployment model will prevail for the foreseeable future.” For now, IT professionals will have to pick and choose from these and potentially more technologies to accelerate their Web services, he says.