Geoff Gerrietts [mailto:[EMAIL PROTECTED] wrote: > I'm working on a very short timetable to accrue a list of viable > technologies for a probable site migration. The story is long, but > Geoff Talvola recommended I post here for insight, so I shall. > > I'm part of a small dev team working on a website with pretty big > traffic. Today, we have zope in front of ILU (yes, ILU) on Python > 1.5.2 as our platform. We have right around 30 servers deployed, over > twenty of them running Zope, and we're having problems keeping up > during peak ours. We expect to be completely overwhelmed by June.
What kind of hits per second are you getting on the Zope servers during a peak hour? > This prompts us to look at switching architectures and technologies. > The guy who is closest in authority and responsibilities to a lead > architect thinks we should scrap the python infrastructure and move to > Java (Tomcat, Struts, JBoss). The development team, aware of what that > means to our development cycle and net productivity, is resisting > that. I'm our best hope for coming up with a working alternative. > > For a long time, I've looked at alternate technologies that we might > be able to use, if a day like this were going to come. Until > yesterday, I did not believe such a day would. Now that it has, I > don't feel like I've come up with solid answers, and very few facts to > support those answers. My personal preference is for the WebKit -- > I've always liked the Servlet/WebObjects architectural style, and I > think WebKit has built that architecture out very well. > > We have a fairly lengthy list of proposed requirements, but the > biggest is that our site is very dynamic on almost every page, and we > have a high volume of traffic that's climbing every day. It's also > worth noting that we plan to stop distributing processes in a > services-based architecture, and plan to distribute only the sessions: > everything else will run on a single box. > > I'm concerned about performance, but less today than yesterday. My > biggest concern today is finding people who have proven technological > solutions in play at high-volume sites. I'd also appreciate a candid > evaluation of the "readiness state" of python technologies, including > WebKit's various components. > > I would also appreciate any pointers to good solutions I might have > overlooked. > > I've examined mod_python and Twisted, and looked at new versions of > Zope, and at WebWare. The former don't appear to do what we want, and > I'm still trying to sort out the details of the latter two. My belief > is that Zope (because DTML is interpreted on the fly) will be too slow > for our needs. If your bottleneck is Zope's CPU usage, then I'm quite confident that Webware will make a huge improvement. It is a much more lightweight architecture, and your servlets are just python code, so they can pretty much run as fast as any Python code, whereas Zope has a lot of machinery layered on top of Python so it ends up being much slower, at least when you're using DTML. Note that PSP (included with Webware) and Cheetah (a separate package) are templating solutions for Webware that compile their templates down to Python code, so if you use them, you'll still get Python's speed. I know I've seen some benchmarks on this list before, but I can't find them now. Is there any reasonable way to search the archives? I'm not sure that anyone has yet implemented a high volume application like this using Webware. There's certainly no reason why it shouldn't work fine. - Geoff ------------------------------------------------------- This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger for complex code. Debugging C/C++ programs can leave you feeling lost and disoriented. TotalView can help you find your way. Available on major UNIX and Linux platforms. Try it free. www.etnus.com _______________________________________________ Webware-discuss mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/webware-discuss
