You are viewing a plain text version of this content. The canonical link for it is here.
Posted to soap-user@xml.apache.org by "R. Mark Volkmann" <vo...@home.com> on 2001/04/27 13:20:58 UTC
roundtrip time > 3 seconds?
I'm excited about SOAP, but I'm also worried that it might not be practical
to write an application that makes many SOAP calls. In my experience, I
never get a remote SOAP call to complete in less than three seconds. It's
much better if the server is on my box, but that's not a realistic example.
Does anyone else get significantly better times for remote SOAP calls?
__ __
/ \/ \ Object Computing, Inc.
\ / ark (314)589-1617 pager
(314)579-0066 OCI
\ / olkmann mark@ociweb.com
\/
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
otpimizations [Re: roundtrip time > 3 seconds?]
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Paul John Gibbons wrote:
>
> Or are there other techniques being used which might limit where these
> higher rates can be achieved?
>
careful optimizing all layers is the key :-)
for us best worked to optimize network (keep-alive helps a lot) and xml
parsing (especially minimize cost of creating/destroying xml parser).
thanks,
alek
--
Aleksander Slominski, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall-Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Paul John Gibbons wrote:
> Do I need to reuse a HTTP 1.1 connection to achieve the ~100 call/second rates?
i am talking about SoapRMI 1.1 and giving example of possible performance
that you could expect on standard machines (PII 233Mhz).
this number is without keep alive - we had to give up on flaky JDK 1.3
client side keep alive (it tents ot get stuck in KeepAlive.close in busy
wait) but when we were using keep-alive we had 300 calls/sec.
thanks,
alek
--
Aleksander Slominski, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall-Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
otpimizations [Re: roundtrip time > 3 seconds?]
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Paul John Gibbons wrote:
>
> Or are there other techniques being used which might limit where these
> higher rates can be achieved?
>
careful optimizing all layers is the key :-)
for us best worked to optimize network (keep-alive helps a lot) and xml
parsing (especially minimize cost of creating/destroying xml parser).
thanks,
alek
--
Aleksander Slominski, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall-Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Paul John Gibbons wrote:
> Do I need to reuse a HTTP 1.1 connection to achieve the ~100 call/second rates?
i am talking about SoapRMI 1.1 and giving example of possible performance
that you could expect on standard machines (PII 233Mhz).
this number is without keep alive - we had to give up on flaky JDK 1.3
client side keep alive (it tents ot get stuck in KeepAlive.close in busy
wait) but when we were using keep-alive we had 300 calls/sec.
thanks,
alek
--
Aleksander Slominski, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall-Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Paul John Gibbons <pa...@smarterville.com>.
Is the bulk of this 1.5-3 second time connection setup?
Do I need to reuse a HTTP 1.1 connection to achieve the ~100 call/second
rates?
Or are there other techniques being used which might limit where these
higher rates can be achieved?
Paul Gibbons
paul@smarterville.com
> > I'm excited about SOAP, but I'm also worried that it might not be
> practical
> > to write an application that makes many SOAP calls. In my experience, I
> > never get a remote SOAP call to complete in less than three seconds.
It's
> > much better if the server is on my box, but that's not a realistic
> example.
> > Does anyone else get significantly better times for remote SOAP calls?
>
> People who claim to be able to get Apache SOAP to do 100 calls/second must
> be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
> return a float) I've tested takes 1.5 seconds to complete...on localhost.
>
> We are extremely disappointed with Apache SOAP's speed.
>
> ---
> Clay
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Help - running the samples
Posted by Eduardo Salvador Godoy Vega <eg...@tekchile.cl>.
Hi,
I just config my Win2K-Apache-Tomcat-Soap system ... but when I want to run
the samples I get some errors...
for examples
D:\Public\TOMCAT\soap\samples\calculator>testit
This test assumes a server URL of
http://localhost:8080/soap/server/rpcrouter
Deploying the calcuator service...
Exception in thread "main" [SOAPException: faultCode=SOAP-ENV:Client; msg=A
'htt
p://schemas.xmlsoap.org/soap/envelope/:Fault' element must contain a:
'faultcode
' element.; targetException=java.lang.IllegalArgumentException: A
'http://schema
s.xmlsoap.org/soap/envelope/:Fault' element must contain a: 'faultcode'
element.
]
at org.apache.soap.rpc.Call.invoke(Call.java:244)
at
org.apache.soap.server.ServiceManagerClient.invokeMethod(ServiceManag
erClient.java:127)
at
org.apache.soap.server.ServiceManagerClient.deploy(ServiceManagerClie
nt.java:140)
at
org.apache.soap.server.ServiceManagerClient.main(ServiceManagerClient
.java, Compiled Code)
.
Verify that its there
Exception in thread "main" [SOAPException: faultCode=SOAP-ENV:Client;
msg=Unable
to resolve namespace URI for 'ns2'.;
targetException=java.lang.IllegalArgumentE
xception: Unable to resolve namespace URI for 'ns2'.]
at org.apache.soap.rpc.Call.invoke(Call.java:244)
at
org.apache.soap.server.ServiceManagerClient.invokeMethod(ServiceManag
erClient.java:127)
at
org.apache.soap.server.ServiceManagerClient.list(ServiceManagerClient
.java:149)
at
org.apache.soap.server.ServiceManagerClient.main(ServiceManagerClient
.java, Compiled Code)
.
Any Idea ... I think that is something "simple" but I can't find the bug...
Thanks
Eduardo.
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Help - running the samples
Posted by Eduardo Salvador Godoy Vega <eg...@tekchile.cl>.
Hi,
I just config my Win2K-Apache-Tomcat-Soap system ... but when I want to run
the samples I get some errors...
for examples
D:\Public\TOMCAT\soap\samples\calculator>testit
This test assumes a server URL of
http://localhost:8080/soap/server/rpcrouter
Deploying the calcuator service...
Exception in thread "main" [SOAPException: faultCode=SOAP-ENV:Client; msg=A
'htt
p://schemas.xmlsoap.org/soap/envelope/:Fault' element must contain a:
'faultcode
' element.; targetException=java.lang.IllegalArgumentException: A
'http://schema
s.xmlsoap.org/soap/envelope/:Fault' element must contain a: 'faultcode'
element.
]
at org.apache.soap.rpc.Call.invoke(Call.java:244)
at
org.apache.soap.server.ServiceManagerClient.invokeMethod(ServiceManag
erClient.java:127)
at
org.apache.soap.server.ServiceManagerClient.deploy(ServiceManagerClie
nt.java:140)
at
org.apache.soap.server.ServiceManagerClient.main(ServiceManagerClient
.java, Compiled Code)
.
Verify that its there
Exception in thread "main" [SOAPException: faultCode=SOAP-ENV:Client;
msg=Unable
to resolve namespace URI for 'ns2'.;
targetException=java.lang.IllegalArgumentE
xception: Unable to resolve namespace URI for 'ns2'.]
at org.apache.soap.rpc.Call.invoke(Call.java:244)
at
org.apache.soap.server.ServiceManagerClient.invokeMethod(ServiceManag
erClient.java:127)
at
org.apache.soap.server.ServiceManagerClient.list(ServiceManagerClient
.java:149)
at
org.apache.soap.server.ServiceManagerClient.main(ServiceManagerClient
.java, Compiled Code)
.
Any Idea ... I think that is something "simple" but I can't find the bug...
Thanks
Eduardo.
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Peter Glynn <gl...@tcd.ie>.
> I've tested takes 1.5 seconds to complete...on localhost.
As you can see from my other mail the lowest result that I got was 61 ms .
So it wasn't a ciggy my buddy beside was smoken?
----- Original Message -----
From: "Clay Luther" <cl...@cisco.com>
To: <so...@xml.apache.org>
Sent: Friday, April 27, 2001 7:32 PM
Subject: Re: roundtrip time > 3 seconds?
> > I'm excited about SOAP, but I'm also worried that it might not be
> practical
> > to write an application that makes many SOAP calls. In my experience, I
> > never get a remote SOAP call to complete in less than three seconds.
It's
> > much better if the server is on my box, but that's not a realistic
> example.
> > Does anyone else get significantly better times for remote SOAP calls?
>
> People who claim to be able to get Apache SOAP to do 100 calls/second must
> be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
> return a float) I've tested takes 1.5 seconds to complete...on localhost.
>
> We are extremely disappointed with Apache SOAP's speed.
>
> ---
> Clay
> Cisco Systems, Inc.
> claycle@cisco.com
> (972) 813-5004
>
> J'ai un regret: je ne regrette rien.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
> For additional commands, email: soap-user-help@xml.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Clay Luther wrote:
> People who claim to be able to get Apache SOAP to do 100 calls/second must
> be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
> return a float) I've tested takes 1.5 seconds to complete...on localhost.
i was talking about SoapRMI and refering to possile optimizations that we
have already done :-)
> We are extremely disappointed with Apache SOAP's speed.
so help them to find what is now working and fix it.
thanks,
alek
--
Aleksander Slominski, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall-Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Clay Luther wrote:
> People who claim to be able to get Apache SOAP to do 100 calls/second must
> be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
> return a float) I've tested takes 1.5 seconds to complete...on localhost.
i was talking about SoapRMI and refering to possile optimizations that we
have already done :-)
> We are extremely disappointed with Apache SOAP's speed.
so help them to find what is now working and fix it.
thanks,
alek
--
Aleksander Slominski, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall-Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Peter Glynn <gl...@tcd.ie>.
> I've tested takes 1.5 seconds to complete...on localhost.
As you can see from my other mail the lowest result that I got was 61 ms .
So it wasn't a ciggy my buddy beside was smoken?
----- Original Message -----
From: "Clay Luther" <cl...@cisco.com>
To: <so...@xml.apache.org>
Sent: Friday, April 27, 2001 7:32 PM
Subject: Re: roundtrip time > 3 seconds?
> > I'm excited about SOAP, but I'm also worried that it might not be
> practical
> > to write an application that makes many SOAP calls. In my experience, I
> > never get a remote SOAP call to complete in less than three seconds.
It's
> > much better if the server is on my box, but that's not a realistic
> example.
> > Does anyone else get significantly better times for remote SOAP calls?
>
> People who claim to be able to get Apache SOAP to do 100 calls/second must
> be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
> return a float) I've tested takes 1.5 seconds to complete...on localhost.
>
> We are extremely disappointed with Apache SOAP's speed.
>
> ---
> Clay
> Cisco Systems, Inc.
> claycle@cisco.com
> (972) 813-5004
>
> J'ai un regret: je ne regrette rien.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
> For additional commands, email: soap-user-help@xml.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Paul John Gibbons <pa...@smarterville.com>.
Is the bulk of this 1.5-3 second time connection setup?
Do I need to reuse a HTTP 1.1 connection to achieve the ~100 call/second
rates?
Or are there other techniques being used which might limit where these
higher rates can be achieved?
Paul Gibbons
paul@smarterville.com
> > I'm excited about SOAP, but I'm also worried that it might not be
> practical
> > to write an application that makes many SOAP calls. In my experience, I
> > never get a remote SOAP call to complete in less than three seconds.
It's
> > much better if the server is on my box, but that's not a realistic
> example.
> > Does anyone else get significantly better times for remote SOAP calls?
>
> People who claim to be able to get Apache SOAP to do 100 calls/second must
> be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
> return a float) I've tested takes 1.5 seconds to complete...on localhost.
>
> We are extremely disappointed with Apache SOAP's speed.
>
> ---
> Clay
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Clay Luther <cl...@cisco.com>.
> I'm excited about SOAP, but I'm also worried that it might not be
practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic
example.
> Does anyone else get significantly better times for remote SOAP calls?
People who claim to be able to get Apache SOAP to do 100 calls/second must
be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
return a float) I've tested takes 1.5 seconds to complete...on localhost.
We are extremely disappointed with Apache SOAP's speed.
---
Clay
Cisco Systems, Inc.
claycle@cisco.com
(972) 813-5004
J'ai un regret: je ne regrette rien.
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by David Wall <dw...@Yozons.com>.
> I'm excited about SOAP, but I'm also worried that it might not be
practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic
example.
> Does anyone else get significantly better times for remote SOAP calls?
Our early tests show lengthy operations as well, but it would appear that
most of the time is in the Call invoke. From what little I can tell, the
time spent in my own service bean is quite small, as expected. But there is
some real nastiness going with the XML stuff, I'd guess.
I should do a more complex testing many invocations to see if that holds.
Maybe there's just some really nasty setup time that I'm not aware of (like
a SecureRandom object being created).
David
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by David Wall <dw...@Yozons.com>.
> however the internet portion of SOAP communication is always going to be
a bottleneck.
It's not the Internet, since these long times can be seen in a LAN
environment. And since we also have the same services available via a web
browser service, you can easily compare the two, where the browser is snappy
and the Call.invoke() is a slug.
David
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by David Wall <dw...@Yozons.com>.
> however the internet portion of SOAP communication is always going to be
a bottleneck.
It's not the Internet, since these long times can be seen in a LAN
environment. And since we also have the same services available via a web
browser service, you can easily compare the two, where the browser is snappy
and the Call.invoke() is a slug.
David
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by "D. Rioux" <dr...@pictorius.com>.
Hi Mark:
Graham Glass contributed an excellent article on a SOAP-specific XML
framework designed to speed up the XML aspects of SOAP, however the
internet portion of SOAP communication is always going to be a
bottleneck. You can check out the article I'm referring to, however:
http://www-106.ibm.com/developerworks/xml/library/x-elexml/index.html
// David Rioux
// drioux@pictorius.com
R. Mark Volkmann wrote:
> I'm excited about SOAP, but I'm also worried that it might not be practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic example.
> Does anyone else get significantly better times for remote SOAP calls?
>
> __ __
> / \/ \ Object Computing, Inc.
> \ / ark (314)589-1617 pager
> (314)579-0066 OCI
> \ / olkmann mark@ociweb.com
> \/
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
> For additional commands, email: soap-user-help@xml.apache.org
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Magnus ?or Torfason wrote:
> > in SoapRMI we routinely get >100 calls per second and in apache soap with
>
> By this you must be talking about throughput. I am running simple samples
> on a single machine and a round-trip time of less than 10 ms is NOT what I
> am getting.
this is round-trip and not throughput (it includes: preparing request, sending
request, deserializing request, processing, serializing reply, sending reply,
deserializing reply). we are right now testing distributed events between
C/C++/Java on w2k/sun/linux and those are numbers for sending events of size
around 1K-2K.
profile your code or just simply try to replace XML parser - i am sure it may
have dramatic effect on your system :-)
thanks,
alek
--
Aleksander Slominski, LH 316, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall - Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Magnus ?or Torfason wrote:
> > in SoapRMI we routinely get >100 calls per second and in apache soap with
>
> By this you must be talking about throughput. I am running simple samples
> on a single machine and a round-trip time of less than 10 ms is NOT what I
> am getting.
this is round-trip and not throughput (it includes: preparing request, sending
request, deserializing request, processing, serializing reply, sending reply,
deserializing reply). we are right now testing distributed events between
C/C++/Java on w2k/sun/linux and those are numbers for sending events of size
around 1K-2K.
profile your code or just simply try to replace XML parser - i am sure it may
have dramatic effect on your system :-)
thanks,
alek
--
Aleksander Slominski, LH 316, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall - Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
RE: roundtrip time > 3 seconds?
Posted by Magnus ?or Torfason <ma...@handtolvur.is>.
> in SoapRMI we routinely get >100 calls per second and in apache soap with
By this you must be talking about throughput. I am running simple samples
on a single machine and a round-trip time of less than 10 ms is NOT what I
am getting.
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
RE: roundtrip time > 3 seconds?
Posted by Magnus ?or Torfason <ma...@handtolvur.is>.
> in SoapRMI we routinely get >100 calls per second and in apache soap with
By this you must be talking about throughput. I am running simple samples
on a single machine and a round-trip time of less than 10 ms is NOT what I
am getting.
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
"R. Mark Volkmann" wrote:
> I'm excited about SOAP, but I'm also worried that it might not be practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic example.
> Does anyone else get significantly better times for remote SOAP calls?
in SoapRMI we routinely get >100 calls per second and in apache soap with
faster parser (try xerces 2 or some small and fast like aelfred 2) you should
have much better results.
thanks,
alek
ps. SoapRMI is open source under BSD-like and you can get it from:
http://www.extreme.indiana.edu/soap/
--
Aleksander Slominski, LH 316, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall - Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
"R. Mark Volkmann" wrote:
> I'm excited about SOAP, but I'm also worried that it might not be practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic example.
> Does anyone else get significantly better times for remote SOAP calls?
in SoapRMI we routinely get >100 calls per second and in apache soap with
faster parser (try xerces 2 or some small and fast like aelfred 2) you should
have much better results.
thanks,
alek
ps. SoapRMI is open source under BSD-like and you can get it from:
http://www.extreme.indiana.edu/soap/
--
Aleksander Slominski, LH 316, IU, http://www.extreme.indiana.edu/~aslom
As I look afar I see neither cherry Nor tinted leaves Just a modest hut
on the coast In the dusk of Autumn nightfall - Fujiwara no Teika(1162-1241)
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by "D. Rioux" <dr...@pictorius.com>.
Hi Mark:
Graham Glass contributed an excellent article on a SOAP-specific XML
framework designed to speed up the XML aspects of SOAP, however the
internet portion of SOAP communication is always going to be a
bottleneck. You can check out the article I'm referring to, however:
http://www-106.ibm.com/developerworks/xml/library/x-elexml/index.html
// David Rioux
// drioux@pictorius.com
R. Mark Volkmann wrote:
> I'm excited about SOAP, but I'm also worried that it might not be practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic example.
> Does anyone else get significantly better times for remote SOAP calls?
>
> __ __
> / \/ \ Object Computing, Inc.
> \ / ark (314)589-1617 pager
> (314)579-0066 OCI
> \ / olkmann mark@ociweb.com
> \/
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
> For additional commands, email: soap-user-help@xml.apache.org
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by Clay Luther <cl...@cisco.com>.
> I'm excited about SOAP, but I'm also worried that it might not be
practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic
example.
> Does anyone else get significantly better times for remote SOAP calls?
People who claim to be able to get Apache SOAP to do 100 calls/second must
be smoking something. The SIMPLEST and FASTEST RPC call (send a string,
return a float) I've tested takes 1.5 seconds to complete...on localhost.
We are extremely disappointed with Apache SOAP's speed.
---
Clay
Cisco Systems, Inc.
claycle@cisco.com
(972) 813-5004
J'ai un regret: je ne regrette rien.
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org
Re: roundtrip time > 3 seconds?
Posted by David Wall <dw...@Yozons.com>.
> I'm excited about SOAP, but I'm also worried that it might not be
practical
> to write an application that makes many SOAP calls. In my experience, I
> never get a remote SOAP call to complete in less than three seconds. It's
> much better if the server is on my box, but that's not a realistic
example.
> Does anyone else get significantly better times for remote SOAP calls?
Our early tests show lengthy operations as well, but it would appear that
most of the time is in the Call invoke. From what little I can tell, the
time spent in my own service bean is quite small, as expected. But there is
some real nastiness going with the XML stuff, I'd guess.
I should do a more complex testing many invocations to see if that holds.
Maybe there's just some really nasty setup time that I'm not aware of (like
a SecureRandom object being created).
David
---------------------------------------------------------------------
To unsubscribe, e-mail: soap-user-unsubscribe@xml.apache.org
For additional commands, email: soap-user-help@xml.apache.org