In a service-oriented architecture, testing is not something to be saved for application runtime. Sure, runtime...
processes and environments must be tested, but a developer greatly benefits from knowing the individual services have been pushed to their limits under a microscope.
One common issue that can hinder service performance is attachments. The transfer of large files has been a problem in IT for many years. Now that the cloud is providing seemingly unlimited on-demand resources, Gartner Research Director, Frank Kenney says it could get worse.
"As we start to use more cloud-based services, the problem is going to exacerbate itself because we're dealing with bigger data, bigger attachments," said Kenney. "But we want the same performance that we've always been able to maintain."
Historically, most applications have been given a more transactional architecture where transfers occurred in small bursts. Now, Kenney says the industry is working to have applications constantly transferring data in real-time.
These sorts of concerns are where SOA testing vendor, Crosscheck Networks has aligned its offerings. The company's SOAPSonar provides tools for functional, performance, compliance and security testing. In the recent version 5 release, the product gained support for WS-Trust and large file streaming tests.
Where FTP has been declining in popularity for file transfers, Web services transfer systems like MIME have become more prevalent, suggested Mamoon Yunus, CEO of Crosscheck. But file sizes are growing faster than many involved with SOA had anticipated.
In addition, enterprises have started to move towards integrating parts of their customers' infrastructures with their own.
"SOA has historically been very internally focused," said Yunus. "But now enterprises are adding their trading partners to their integration profiles. The first thing of value that they're hitting is how you trust who's coming into your Web services."
The more external access granted to one's system, the more possibilities for attack vectors and mutated data injections.
Dig Deeper on APM and log analysis