[From nobody Wed Jul 13 02:52:25 2011
Message-ID: <3DA6E634.409@jhu.edu>
Date: Fri, 11 Oct 2002 10:54:44 -0400
From: Ryan Caudy <caudy@jhu.edu>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
	rv:1.1) Gecko/20020826
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Jerry Asher <jerry-spread@theashergroup.com>
Subject: Re: [Spread-users] Spread newbie -- spread application questions....
References: <5.1.1.6.0.20021011023508.036eea08@theashergroup.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

My responses are interleaved.

--Ryan Caudy

Jerry Asher wrote:

>I am building an application involving several hundred servers all behind our firewall.  I would like to send various heartbeat and status messages around, and I am thinking that I could use a multicast application and/or Spread.
>
>I have read the spread faq and the first 15 pages of the spread user guide.
>
>Did I interpret a statement correctly that Spread isn't really applicable for a system of several hundred machines?
>  
>
Whether Spread is able to do what you like on a system of several 
hundred machines depends on the layout of your spread network, and the 
decisions you make.  If you tell us a little more about your situation, 
this list can probably help you out.

>Also, another developer tells me that when he looked at Spread he discovered that many of Spread's reliability benefits derive from Spread writing messages to disks.  Is that true?  Does spread involve the disk (we would prefer to find something that uses only memory....)?
>  
>
Spread does NOT write any data to disk, besides logging information if 
you request it to do so.  Its reliable messaging service is based on 
communication between Spread daemons that determines when 
retransmissions are necessary.  None of Spread's other services require 
writing to stable storage, either.

>Finally, is there a specific format for spread messages, or do we layer our own protocol over spread?  I would basically like to tunnel http requests over the spread "backbone".  Why? Because my favorite swiss army knife is aolserver and that is a highly programmable and modifiable webserver.  So how hard would it be to tunnel http gets or posts across spread?
>  
>
Spread messages have no specific format at the application level.  The 
SP_muliticast and other send routines do not match the API of any other 
protocols I know of, but this is largely because these functions have 
specific fields related to the services Spread provides.  The contents 
of the "mess_type" and "mess_len" fields, as well as the "mess" 
character array are up to you, though I would make sure that mess_len 
actually specifies the length of the data in the character array that 
you would like to transmit.

Spread does not provide a stream abstraction, i.e. like TCP, which you 
might want to simplify tunneling.  However, it shouldn't be too 
difficult to write code that provides that abstraction to your applications.

Good luck!

>Thank you,
>
>
>Jerry Asher 
>
>
>_______________________________________________
>Spread-users mailing list
>Spread-users@lists.spread.org
>http://lists.spread.org/mailman/listinfo/spread-users
>
>  
>

]