You are viewing a plain text version of this content. The canonical link for it is here.
Posted to soap-user@ws.apache.org by Christian Aberger <xm...@Aberger.at> on 2001/03/18 10:08:54 UTC
DCOM Http Tunneling versus SOAP
Hi,
I have a question. As I was adviced in this list, I had a look into
Micro$oft-SOAP toolkit 2.0 beta. It is easily installed and the samples seem
to work. Well, nice.
Now the question. All that the toolkit does is to support that client side
COM(TM)-objects can call server side COM(TM) objects. The inner transport
layer propably is SOAP, but neither client nor server ever see it, so why is
it called SOAP toolkit ? Maybe because it is slow ("Slow Object Access
Protocol")? ;-)
So what is the difference to DCOM(TM) http-tunneling that we have and use
since NT(TM)-SP4 was released ? Now anyway COM(TM) is abondoned in favour of
.net(TM), so why write still and code for COM(TM) internet services when the
.net(TM) release is near?
History:
1) There was about 20 years ago RPC on unix (remote procedure calls, open
standard) that describes how to serialize data structures in a a standard
binary form through a wire. Then long time later there was M$-RPC (also
called "DCE"-RPC) that did the same but was incompatible.
2) OLE2(TM), "object linkink and embedding", renamed to COM(TM).
3) DCOM as an incompatible and more complicated CORBA over DCE-RPC, but can
be called also with Vi$ual Basic(TM).
4) Security problems on the internet --> firewalls established, RPC ports
(and COM(TM) Ports) not accessible over the internet ("internet" is not yet
a trademark I think ;-/), so DCOM(TM) does not work over the internet any
more.
5) DCOM(TM)-http - tunneling redirects RPC through port 80 and deploys the
calls on the server again, same marshalling back, often works for some time
(if HTTP_KEEP_ALIVE works and the server is rebooted regularily because of
memory leaks etc).
6) 2000: SOAP W3C recommendation: similar to (1) describes how to marshal
data structures now not in binary. Nice, an open standard, although it does
this in ASCII form through a wire (well, we have faster internet lines now,
so we can loose some speed, never mind). And, alas, no need to fix previous
point any more.
7) 2001: Micro$oft encapsulates the open standard to make it accessible only
under COM(TM)/windoze (until the apache module in C++ is done). With Apache
SOAP/Tomcat/Sun(TM)-Java(TM) it is currently already possible under *ux.
So can anybody please explain me why all that histery with SOAP ?
wfR Christian Aberger
The trademarks mentioned in this article are trademarks of the according
companies. As far as I am informed "internet" and "air" are not yet a
trademark. If so it was not my intention to use these terms in this mail.
;-)
Re: DCOM Http Tunneling versus SOAP
Posted by Russell Beattie <rb...@mail.com>.
Hi Christian,
Yes, Microsoft generally sucks. But in this case, it was Microsoft,
DevelopMentor and Userland who came up with the SOAP spec and then later on
IBM and others jumped on the bandwagon because it was a good idea. The guys
in Seattle do have them once in a while. Remember also that the W3C is not
an open organization out the help the common good, it's membership is
comprised only of paying companies like Microsoft and others. You and I for
example, can't submit a proposal for a new internet spec, only it's members
are allowed to do this.
So the question is, if Microsoft has this history of being proprietary, why
did they come up with SOAP - especially since as you point out DCOM can work
of HTTP? And why are they enabling Office apps to save everything in XML
format since that will only enable competition. I assume that Microsoft is
up to something, but what? Bill's a lot smarter guy than me (and obviously a
lot richer), so I'm sure I'll figure it out only after it's too late...
For more info about how SOAP was created, check out Dave's history of SOAP
from Userland.
http://www.xmlrpc.com/stories/storyReader$555
And check out the authors for the W3C submission.
http://www.w3.org/TR/2000/NOTE-SOAP-20000508/
-Russ
----- Original Message -----
From: "Christian Aberger" <xm...@Aberger.at>
To: <so...@xml.apache.org>
Sent: Sunday, March 18, 2001 10:08 AM
Subject: DCOM Http Tunneling versus SOAP
> Hi,
>
> I have a question. As I was adviced in this list, I had a look into
> Micro$oft-SOAP toolkit 2.0 beta. It is easily installed and the samples
seem
> to work. Well, nice.
>
> Now the question. All that the toolkit does is to support that client side
> COM(TM)-objects can call server side COM(TM) objects. The inner transport
> layer propably is SOAP, but neither client nor server ever see it, so why
is
> it called SOAP toolkit ? Maybe because it is slow ("Slow Object Access
> Protocol")? ;-)
>
> So what is the difference to DCOM(TM) http-tunneling that we have and use
> since NT(TM)-SP4 was released ? Now anyway COM(TM) is abondoned in favour
of
> .net(TM), so why write still and code for COM(TM) internet services when
the
> .net(TM) release is near?
>
> History:
> 1) There was about 20 years ago RPC on unix (remote procedure calls, open
> standard) that describes how to serialize data structures in a a standard
> binary form through a wire. Then long time later there was M$-RPC (also
> called "DCE"-RPC) that did the same but was incompatible.
> 2) OLE2(TM), "object linkink and embedding", renamed to COM(TM).
> 3) DCOM as an incompatible and more complicated CORBA over DCE-RPC, but
can
> be called also with Vi$ual Basic(TM).
> 4) Security problems on the internet --> firewalls established, RPC ports
> (and COM(TM) Ports) not accessible over the internet ("internet" is not
yet
> a trademark I think ;-/), so DCOM(TM) does not work over the internet any
> more.
> 5) DCOM(TM)-http - tunneling redirects RPC through port 80 and deploys the
> calls on the server again, same marshalling back, often works for some
time
> (if HTTP_KEEP_ALIVE works and the server is rebooted regularily because of
> memory leaks etc).
>
> 6) 2000: SOAP W3C recommendation: similar to (1) describes how to marshal
> data structures now not in binary. Nice, an open standard, although it
does
> this in ASCII form through a wire (well, we have faster internet lines
now,
> so we can loose some speed, never mind). And, alas, no need to fix
previous
> point any more.
>
> 7) 2001: Micro$oft encapsulates the open standard to make it accessible
only
> under COM(TM)/windoze (until the apache module in C++ is done). With
Apache
> SOAP/Tomcat/Sun(TM)-Java(TM) it is currently already possible under *ux.
>
> So can anybody please explain me why all that histery with SOAP ?
>
> wfR Christian Aberger
>
> The trademarks mentioned in this article are trademarks of the according
> companies. As far as I am informed "internet" and "air" are not yet a
> trademark. If so it was not my intention to use these terms in this mail.
> ;-)
Re: DCOM Http Tunneling versus SOAP
Posted by Russell Beattie <rb...@mail.com>.
Hi Christian,
Yes, Microsoft generally sucks. But in this case, it was Microsoft,
DevelopMentor and Userland who came up with the SOAP spec and then later on
IBM and others jumped on the bandwagon because it was a good idea. The guys
in Seattle do have them once in a while. Remember also that the W3C is not
an open organization out the help the common good, it's membership is
comprised only of paying companies like Microsoft and others. You and I for
example, can't submit a proposal for a new internet spec, only it's members
are allowed to do this.
So the question is, if Microsoft has this history of being proprietary, why
did they come up with SOAP - especially since as you point out DCOM can work
of HTTP? And why are they enabling Office apps to save everything in XML
format since that will only enable competition. I assume that Microsoft is
up to something, but what? Bill's a lot smarter guy than me (and obviously a
lot richer), so I'm sure I'll figure it out only after it's too late...
For more info about how SOAP was created, check out Dave's history of SOAP
from Userland.
http://www.xmlrpc.com/stories/storyReader$555
And check out the authors for the W3C submission.
http://www.w3.org/TR/2000/NOTE-SOAP-20000508/
-Russ
----- Original Message -----
From: "Christian Aberger" <xm...@Aberger.at>
To: <so...@xml.apache.org>
Sent: Sunday, March 18, 2001 10:08 AM
Subject: DCOM Http Tunneling versus SOAP
> Hi,
>
> I have a question. As I was adviced in this list, I had a look into
> Micro$oft-SOAP toolkit 2.0 beta. It is easily installed and the samples
seem
> to work. Well, nice.
>
> Now the question. All that the toolkit does is to support that client side
> COM(TM)-objects can call server side COM(TM) objects. The inner transport
> layer propably is SOAP, but neither client nor server ever see it, so why
is
> it called SOAP toolkit ? Maybe because it is slow ("Slow Object Access
> Protocol")? ;-)
>
> So what is the difference to DCOM(TM) http-tunneling that we have and use
> since NT(TM)-SP4 was released ? Now anyway COM(TM) is abondoned in favour
of
> .net(TM), so why write still and code for COM(TM) internet services when
the
> .net(TM) release is near?
>
> History:
> 1) There was about 20 years ago RPC on unix (remote procedure calls, open
> standard) that describes how to serialize data structures in a a standard
> binary form through a wire. Then long time later there was M$-RPC (also
> called "DCE"-RPC) that did the same but was incompatible.
> 2) OLE2(TM), "object linkink and embedding", renamed to COM(TM).
> 3) DCOM as an incompatible and more complicated CORBA over DCE-RPC, but
can
> be called also with Vi$ual Basic(TM).
> 4) Security problems on the internet --> firewalls established, RPC ports
> (and COM(TM) Ports) not accessible over the internet ("internet" is not
yet
> a trademark I think ;-/), so DCOM(TM) does not work over the internet any
> more.
> 5) DCOM(TM)-http - tunneling redirects RPC through port 80 and deploys the
> calls on the server again, same marshalling back, often works for some
time
> (if HTTP_KEEP_ALIVE works and the server is rebooted regularily because of
> memory leaks etc).
>
> 6) 2000: SOAP W3C recommendation: similar to (1) describes how to marshal
> data structures now not in binary. Nice, an open standard, although it
does
> this in ASCII form through a wire (well, we have faster internet lines
now,
> so we can loose some speed, never mind). And, alas, no need to fix
previous
> point any more.
>
> 7) 2001: Micro$oft encapsulates the open standard to make it accessible
only
> under COM(TM)/windoze (until the apache module in C++ is done). With
Apache
> SOAP/Tomcat/Sun(TM)-Java(TM) it is currently already possible under *ux.
>
> So can anybody please explain me why all that histery with SOAP ?
>
> wfR Christian Aberger
>
> The trademarks mentioned in this article are trademarks of the according
> companies. As far as I am informed "internet" and "air" are not yet a
> trademark. If so it was not my intention to use these terms in this mail.
> ;-)