|
Nathan, from our discussions in the past, I have to believe that an RPG approach, especially your implementation, is the fastest solution. From a common sense perspective (something that seems to be disturbingly lacking in our industry of late), this makes sense: RPG is "closer to the bare metal" than Java. I prefer servlets because Java makes it easier to implement the OO techniques I use in UI design, but Java isn't as fast as RPG. I don't think I'd be able to write my display file emulation in RPG, or at least not as quickly and completely as I have been able to implement it in Java, but I pay a certain performance penalty for that capability. Of course, that's why I have always proposed a Java UI combined with RPG busines logic. Unfortunately, nobody is really trying to get serious performance numbers. I think we as a community need to try to come up with a set of truly representative benchmarks. A "web application" is not a simple, homogenous concept. There are widely varying requirements. For example, a pure query on keyed data (such as an order status inquiry) is a different beast than a data mining application (such as a product search based on keywords, with user-definable sort criteria). A data-entry application is a different category altogether. I'm reasonably certain such tests would indicate that different techniques are more or less appropriate for each of these implementation. Without true apples to apples comparisons, which take time and resources, it's unlikely that a "number of transactions per second per CPW" is going to give you any sort of realistic frame of reference. Joe > -----Original Message----- > From: Nathan M. Andelin > > Quote from page 102: "The data in the Table 6.1 assumes that a > small amount > of data is being served (say 100 > bytes)." > > How relevant are these benchmarks? 100 bytes? Apparently no database > access? Apparently the response is produced entirely from in-line output > statements? > > Properly structured Java applications normally involve a Servlet, a Java > Bean, and a JSP, and frequently a Session, to produce a response. > > Most Net.Data macros evoke calls to the SQL language environment. > > The referenced comparisons seem to be almost completely divorced from > real-life.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.