Re: [Fwd: Re: Mainframe not a good architecture for interactive wasRe: What is the future of COBOL? Answer: Irrelevant???]]
From: Michael Wojcik (mwojcik_at_newsguy.com)
Date: 01/28/04
- Next message: docdwarf_at_panix.com: "Re: [Fwd: Re: Mainframe not a good architecture for interactive was Re: What is the future of COBOL? Answer: Irrelevant???]]"
- Previous message: Peter E.C. Dashwood: "Re: OO? Oh oh!"
- In reply to: Hugh Candlin: "Re: [Fwd: Re: Mainframe not a good architecture for interactive wasRe: What is the future of COBOL? Answer: Irrelevant???]]"
- Next in thread: Peter E.C. Dashwood: "Re: [Fwd: Re: Mainframe not a good architecture for interactive was Re: What is the future of COBOL? Answer: Irrelevant???]]"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 28 Jan 2004 21:23:17 GMT
In article <RwwRb.120657$6y6.2389843@bgtnsc05-news.ops.worldnet.att.net>, "Hugh Candlin" <no@spam.com> writes:
>
> John S. Giltner, Jr. <giltjr@earthlink.net> wrote in message news:zTlRb.146$zG1.71@newsread3.news.atl.earthlink.net...
> >
> > Yes, HTTP as a defined protocol can create as many connection as allowed
> > However, most modern web browser limit the number of connections created
> > to render a single web page to 5.
> Microsoft Knowledge Base Article - 183110
> INFO: WinInet Limits Connections Per Server
> http://support.microsoft.com/default.aspx?scid=kb;en-us;183110
We're off-topic (as usual), but here are a couple of brief comments
about that article:
1. It's out of date. RFC 2068 is no longer the HTTP/1.1 specification;
RFC 2616 superceded it. However, the relevant material in this case
appears to be unchanged.
2. It's misleading if read over-broadly. It's about WinInet, not HTTP
clients in general. RFC 2616 says:
Clients that use persistent connections SHOULD limit the number of
simultaneous connections that they maintain to a given server. A
single-user client SHOULD NOT maintain more than 2 connections with
any server or proxy. (8.1.4)
Points to note:
- This is a SHOULD requirement. It's mandatory for full compliance,
but not for conditional compliance.
- It applies only to clients that use persistent connections. That's
probably not the case in the example John gave, since persistent
connections are generally combined with pipelining, and a pipelining
client doesn't want to open multiple connections to the same server.
- It applies only to single-user clients.
- There are other recommendations for proxies. If your HTTP server is
serving a fully-compliant HTTP/1.1 proxy, it might well have more than
2, or 5, connections.
HTTP/1.1 is a fairly complex protocol - it's not LU6.2, but it's not
SMTP either. It's often not the best choice for an example because
it's hard to remember everything in the spec, existing practice is
frequently non-compliant, and many people are confused about what the
spec actually requires.
-- Michael Wojcik michael.wojcik@microfocus.com HTML is as readable as C. You can take this either way. -- Charlie Gibbs
- Next message: docdwarf_at_panix.com: "Re: [Fwd: Re: Mainframe not a good architecture for interactive was Re: What is the future of COBOL? Answer: Irrelevant???]]"
- Previous message: Peter E.C. Dashwood: "Re: OO? Oh oh!"
- In reply to: Hugh Candlin: "Re: [Fwd: Re: Mainframe not a good architecture for interactive wasRe: What is the future of COBOL? Answer: Irrelevant???]]"
- Next in thread: Peter E.C. Dashwood: "Re: [Fwd: Re: Mainframe not a good architecture for interactive was Re: What is the future of COBOL? Answer: Irrelevant???]]"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|