ASP.NET vs J2EE - Scalability and Speed

Posted: 2004-12-21 19:44:00

java-programmer, ASP.NET vs J2EE - Scalability and Speed Regarding this pieces of information:
65% of the Fortune 500 companies
80% of the top US healthcare companies
75% of commercial banks worldwide
90% of the top commercial banks in the US
67% of the world's largest banks use IBM messaging servers
15 of the top Wall Street brokerage firms
7 of the 8 largest US telecommunications companies

That's true. But what I can understand about that figure was -
Those businesses started using Internet and Computer Networking
technologies times ago. It happens before Microsoft started to get
strong. I am working in a pharmaceutical company in US and we are using
lots of IBM products. We are trying very hard to stop using them and
turn our techology to Microsoft. Off course, personnally - I hate
microsoft - I dont' feel fun when I use any one of micrsoft product.
But in the business term - Microsoft products are qualified. Cheap to
support and run - some tools are never over kill such as SQL server. We
are using Oracle but I found only SQL server that just more than enough
and a lot cheaper to run (cost of DBA and Programmer, etc).

Webspere? Naaa .... we have been using Webspere when we started having
our own Web site and Intranet sites. Then we changed to Coldfusion in
the last 5 years. I found those suck! Cost too much money - when
Microsoft IIS5 or 6 are free also .NET - so, why not!!!!

Sure - a lot of Banks use J2EE - same reason! They rally want to stop
using it but they cannot because everybody started with J2EE and know
J2EE better than anything else!!!

Sad but true!

Posted: 2004-12-22 7:01:00

java-programmer >> ASP.NET vs J2EE - Scalability and Speed
asj wrote:
> "Jeremy Haile" <email***@***.com> wrote in message
news:<bj380n$9gv$email***@***.com>...
> > I work for a software company and it is my decision on whether to
go with
> > J2EE or ASP.NET on an upcoming e-commerce website project. I favor
J2EE
> > mostly because it is what I am most experienced with. However, I
feel that
> > I am going to be challenged on the merits of J2EE vs. ASP.NET.
> >
> > Does anyone in this forum know of any unbiased benchmarks comparing
the
> > scalability or speed of ASP.NET and J2EE? Also, are there any good
examples
> > of large production e-commerce sites that run off of J2EE. What
about
> > statistics of ASP.NET commercial sites vs. J2EE? Finally, I know
the .NET
> > vs J2EE debate has been done a million times, but are there any
good
> > resources that I can use as evidence that J2EE at least stacks up
comparably
> > (or is better) to .NET as a solution for a scalable, e-commerce web
> > application?
>
>
> i believe Microsoft, as part of its license, has prevented the
> publishing of any real independent benchmarks....btw, the selection
of
> a technology is very much dependent on factors beyond just
scalability
> or performance...if you are a mainly microsoft shop and do not want
to
> take an effort to convert to Java, then asp will be the better
> choice.
>

Microsoft has tried to prevent the publishing of .net benchamrks by
trying to enforce the the terms of the MSDN EULA. The MSDN EULA
requires the prior consent before publishing .net benchmark. There has
been a number of benchmarks that show Java something like 10 or 20
times faster than .net. I you do some research you will find these
benchmarks. I don't have them now, but I will post them when I have
more time to go back and find them.

> on the other hand, if the ff holds true, then j2ee would be a better
> choice:
> (1) you will be running on a platform other than windows now or in
> future;
> (2) you want a solution that will scale better;
> (3) if you want a solution that will allow you to project your java
> capabilities easily to areas beyond the server such as cellphones,
> pdas,
>
> for more technical merits of j2ee over .net, one resource is carlos
> perez' manageability site, with his 101 reasons to use j2ee over
.net:
>
> here's a recent independent (pc mag) review of the different j2ee app
> servers and windows server, which has some good points about the
> problems inherent in using .net:
>
> http://www.pcmag.com/article2/0,4149,1218682,00.asp
>
> quotes from this article:
>
> -----------------------------------
>
> - Choosing Microsoft means giving up cross-vendor compatibility, and
> you'll have to forget about running servers on Linux.
>
> - Unlike solutions from BEA and IBM, .NET doesn't offer the
> reliability and security needed for enterprise Web service deployment
>
> - The .NET path offers fewer options in building business logic and
> database components. Microsoft has no official blueprint for business
> objects comparable to Enterprise JavaBeans (EJB), though it
recommends
> best practices on the Web (www .microsoft.com/resources/practices).
> The .NET developers have to devise their own component models based
on
> these practices, while a J2EE developer just needs to run a wizard to
> get EJB.
>
> - While it is easier to field a .NET solution quickly, the rigidity
of
> the EJB specification ensures that large-scale development projects
> can enforce best coding practices and design rules. But this
> stringency comes at the cost of more difficulty for the average
> developer to master. .NET makes development and deployment very easy,
> yet it requires extensive customization if you want to build a
> mission-critical environment in which reliability, transaction
> integrity, and message queue management are essential.
>
> - If your company chooses to develop mission-critical applications in
> .NET, you'll find yourself at Microsoft's mercy regarding upgrades,
> licenses, and support.
>
> ---------------------------------------------------
>
>
> here's something that will impress anyone...if you visit the websites
> of BEA or IBM, you'll find some extremely good examples of very large
> companies using their j2ee app servers. BEA WEblogic's J2EE app
server
> is used in 30,000 customers worldwide, while IBM's J2EE app server
> Websphere provides service to:
>
> 65% of the Fortune 500 companies
> 80% of the top US healthcare companies
> 75% of commercial banks worldwide
> 90% of the top commercial banks in the US
> 67% of the world's largest banks use IBM messaging servers
> 15 of the top Wall Street brokerage firms
> 7 of the 8 largest US telecommunications companies
>
>
> J2EE yanking customers from .net:
> eBay (now THAT'S and extremely large use of J2EE):
> http://www.javalobby.org/thread.jsp?forum=61&thread=8534
>
> Cerner and others:
>
http://groups.google.com/groups?q=cerner+j2ee+.net+asj&hl=en&lr=&ie=UTF-8&selm=3EDE5601.6DBF%40xxx.com&rnum=1

java-programmer >> ASP.NET vs J2EE - Scalability and Speed >> i believe Microsoft, as part of its license, has prevented the
>> publishing of any real independent benchmarks....btw, the selection
> of
>> a technology is very much dependent on factors beyond just
> scalability

>
>
> Microsoft has tried to prevent the publishing of .net benchamrks by
> trying to enforce the the terms of the MSDN EULA. The MSDN EULA
> requires the prior consent before publishing .net benchmark.
>
>

Microsoft troll here . . .

Ahm, yes, it is true that the .NET EULA has a clause in it that says, you
may not publish benchmarks with out Microsoft's consent. Basically, this
wording was put in there to allow Microsoft to respond to any benchmarks
that are published. Microsoft has never (ever ever) declined permission to
anyone, when they requested permission to publish a .NET benchmark they
themselves conducted.

I don't think the MSDN EULA applies in this case.

> There has
> been a number of benchmarks that show Java something like 10 or 20
> times faster than .net. I you do some research you will find these
> benchmarks. I don't have them now, but I will post them when I have
> more time to go back and find them.

I would be very surprised to see results like this. In my experience, the
performance of modern JVMs (like 1.4.2 and above) is very comparable to the
performance of the .NET CLR. If you look you can find numerous smaller
"microbenchmarks", conducted by unbiased third parties, often from academia,
that document such things. They analyze floating point calcs, integer math,
loop control, string manipulation, and some higher level functions such as
XML parsing and transforming, database access, transactional performance,
file and socket I/O, encryption, and so on.

In general, for well-designed systems, Java and .NET will perform similarly.
There will be edge cases where one or the other will be noticeably faster in
some area (let's say, text processing, or forward-only XML parsing), but
those areas don't favor, in general, one over the other, but rather
sometimes Java and sometimes .NET.

We (Microsoft) generally encourage people to conduct their own comparisons,
evaluating performance among the many other criteria that may be relevant
for any particular situation. Among those might be - multi OS capability
(but then again, maybe not !!), developer productivity, capital cost,
security, reliability, availability of developer talent, and ability to
interoperate with existing systems. and I'm sure there are many more
criteria you could consider. To encourage this, we publish source code and
test scripts for benchmarks, including both Java code and .NET code, and
tell people to run their own. It's not drop-dead easy to run your own
tests, but it is direct and irrefutable. Go to
www.gotdotnet.com/team/compare to check out the possibilities there.

The one area where .NET clearly lags in performance ... is on, for example,
AIX, or Solaris. ;)

--
Dino Chiesa
Microsoft Developer Division
d i n o c h @ OmitThis . m i c r o s o f t . c o m

java-programmer >> ASP.NET vs J2EE - Scalability and Speed email***@***.com wrote:
> asj wrote:
> Microsoft has tried to prevent the publishing of .net benchamrks by
> trying to enforce the the terms of the MSDN EULA. The MSDN EULA
> requires the prior consent before publishing .net benchmark. There
has
> been a number of benchmarks that show Java something like 10 or 20
> times faster than .net. I you do some research you will find these
> benchmarks. I don't have them now, but I will post them when I have
> more time to go back and find them.

I'm not sure whether Java or .NET is faster at computation, but it's
very obvious that ngen-ed .NET apps start much faster, and my ASP.NET
worker had never used more than 60MB RAM (with 3-5 web apps).