You are forgetting that if a request takes half the time to complete, it needs less cpu time and you can double the concurrent requests per server.
Evert Dave Wolf wrote: > I simply have to disagree here. We can demonstrate several in > production applications which we have developed using SOAP XML > WebServices and they perform like a champ. One of them was the runner > up for last years MAX award. The majority of the applications we > develop use this architecture and to date not a single time has a > client nor a user complained about the performance of runtime data > services based on SOAP. > > There are a few false rumors that continue to creep up in the Flex > community about the performance issues around SOAP. > > There are benchmarks which show that AMF can be drastically faster > than a SOAP call for the same data. Sometimes even 100% faster. Yup > that's true there are. But you have to peel away the layers of the > onion to see the reality. Statistics can be misleading. For > instance, if AMF is 300 milliseconds and SOAP is 600 milliseconds the > 100% difference isnt even relative. How many people do you know who > can even see 1/3 of a seconds difference? In the end raw marshalling > isnt the issue, it is the user and their experience. Flex2 made > DRASTIC improvements it the performance of XML parsing and in our own > benchmarks the delta between the two services choices is often as low > as 10%. > > Of a much greater impact that the marshalling time is the UI > "shredding" and binding of the data. Most badly performing RIA's > suffer from data being returned from the back-end in a format that > holds no fidelity with the RIA. This requires the RIA to tear apart > the returned structural data and place it into its own structures and > objects and bind those to UI controls. Developing your user > experience in a front-to-back approach which assures great fidelity > between the data formats of the tiers can account for an order of > magnitude performance increase. That is the kind of performance > increase users will actually experience. > > There are many other very smart things you can do like extending > existing controls to do streaming rendering of data to provide the > perception of speed, server side paging, caching, etc. > > In the end perception is reality. All that matters from the UI > perspective is the experience that the user has. Worring about 300 > milliseconds is like trying to debate the number of angels that could > dance on the end of a pin. If the user can't see them, it doesn't > matter how many there are. > > The running rumor that you simply cannot develop first class RIAs in > Flex using a SOAP web services back-end is simply not accurate, and we > have the apps in production with our clients to prove it. > > > -- Flexcoders Mailing List FAQ: http://groups.yahoo.com/group/flexcoders/files/flexcodersFAQ.txt Search Archives: http://www.mail-archive.com/flexcoders%40yahoogroups.com Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/flexcoders/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
