Most services (Web services, XML-RCP, URL APIs, etc) take
a considerable amount of overhead to process a request. For example, in Web services, a
decent amount of XML is being manipulated; this in turn creates a considerable amount
of garbage. Consider that overhead on top of the need to actually service the request by
hitting data sources or performing calculations. BluePages receives approximately 3.2 million
service calls per day and supports a response time of a few milliseconds for service
exploiters. To achieve this level of efficiency, caching was employed to enable smarter
transaction processing.
The key to smarter transaction processing is providing a proxy between the transaction
request and the service logic, or process. The proxy determines if the request should pass
into the process stage or if the request for whatever reason deserves a cached, static, or
predefined response.
In the BluePages architecture, this ???service cache layer??? pre-fetches basic information related
to the service and decides if it has enough information to respond. More specifically,
BluePages offers a URL API to access a user??™s personal photo. The cache layer stores all
Figure 4. Using caching to avoid unnecessary transactions
Goodman, Inampud , & Doran
Copyright ?© 2007, Idea Group Inc.
Pages:
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317