My interpretation is that a transport protocol has no semantic bias -- it just carries bits. A transfer protocol differentiates betwen application-level control vs. data, thus there is a differentiation in what the bits do. It introduces a semantic split between intent ("method") and content.
Someone mentioned Robert C. Martin's OO design principles. If you applied the Interface Seggregation Principle (ISP), Stable Dependencies Principle (SDP) to protocol design, it would lead to a set of small, simple, and general interfaces. Which is really how most Internet transfer protocols are designed. For more specific and complex interaction, one could foresee some kind of policy & contract negotation (and learn a lot from how Telnet did this).
This seems somewhat contrary to the popular way that web services are defined in WSDL. Which to me is because WSDL looks at things 'programmer first', i.e. how to I bind my code to my network protocol. This is different from 'network first', i.e. how do I interoperate and stay flexible.
Cheers
Stu
----- Original Message ----
From: Sanjiva Weerawarana <[EMAIL PROTECTED]>
To: [email protected]
Sent: Monday, May 15, 2006 1:50:31 PM
Subject: Re: [service-orientated-architecture] Re: transfer vs. transport protocols
On Mon, 2006-05-15 at 14:07 +0000, patrickdlogan wrote:
>
> These are two different things for different purposes, the application
> performance and the network performance.
>
> HTTP is at the application layer and can send some information before
> others assuming the receiving application can make good use of it
> without waiting for all of it.
>
> TCP is at the transport layer and can send some information in
> separate packets assuming the overall network will work better with a
> steady flow of similar, small-sized units of information. Then it will
> reassemble them on the other end so you don't have to.
Yes I know all that ;-). I was trying to nail down what the fundamental
difference between a transfer and transport protocol is. The best I've
seen so far is "its the layers in the OSI ref model stupid" .. which is
hardly satisfying .. at least to me.
Anyway, thanks for the responses; I didn't really want to have a long
thread on this- I was looking for an enlightened one to give a solid,
simple definition but that does not appear to be forthcoming.
Sanjiva.
YAHOO! GROUPS LINKS
Someone mentioned Robert C. Martin's OO design principles. If you applied the Interface Seggregation Principle (ISP), Stable Dependencies Principle (SDP) to protocol design, it would lead to a set of small, simple, and general interfaces. Which is really how most Internet transfer protocols are designed. For more specific and complex interaction, one could foresee some kind of policy & contract negotation (and learn a lot from how Telnet did this).
This seems somewhat contrary to the popular way that web services are defined in WSDL. Which to me is because WSDL looks at things 'programmer first', i.e. how to I bind my code to my network protocol. This is different from 'network first', i.e. how do I interoperate and stay flexible.
Cheers
Stu
----- Original Message ----
From: Sanjiva Weerawarana <[EMAIL PROTECTED]>
To: [email protected]
Sent: Monday, May 15, 2006 1:50:31 PM
Subject: Re: [service-orientated-architecture] Re: transfer vs. transport protocols
On Mon, 2006-05-15 at 14:07 +0000, patrickdlogan wrote:
>
> These are two different things for different purposes, the application
> performance and the network performance.
>
> HTTP is at the application layer and can send some information before
> others assuming the receiving application can make good use of it
> without waiting for all of it.
>
> TCP is at the transport layer and can send some information in
> separate packets assuming the overall network will work better with a
> steady flow of similar, small-sized units of information. Then it will
> reassemble them on the other end so you don't have to.
Yes I know all that ;-). I was trying to nail down what the fundamental
difference between a transfer and transport protocol is. The best I've
seen so far is "its the layers in the OSI ref model stupid" .. which is
hardly satisfying .. at least to me.
Anyway, thanks for the responses; I didn't really want to have a long
thread on this- I was looking for an enlightened one to give a solid,
simple definition but that does not appear to be forthcoming.
Sanjiva.
YAHOO! GROUPS LINKS
- Visit your group "service-orientated-architecture" on the web.
- To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
- Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
SPONSORED LINKS
| Computer software | Computer aided design software | Computer job |
| Soa | Service-oriented architecture |
YAHOO! GROUPS LINKS
- Visit your group "service-orientated-architecture" on the web.
- To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
- Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
