[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::fddi

Title:FDDI - The Next Generation
Moderator:NETCAD::STEFANI
Created:Thu Apr 27 1989
Last Modified:Thu Jun 05 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2259
Total number of notes:8590

760.0. "REAL TIME FDDI design" by TOOIS1::MIRGHANE () Tue Oct 27 1992 08:40

    I am working on the design of a real time complex system using some 40
    computers linked with FDDI
    The REAL TIME constraints requires using subnetworks separated by
    computers and or FDDI-FDDI bridges (DECnis bridges)
    From what I understand DECconcentrators do not filter trafic from
    So we have come to the following architecture:


    Can someone help me on the 2 following questions:
    1/ Does it work as designed?
    2/ What is the best way to implement the 4 miles distance? between
       2 linked DECconcentrators or a DECconcentrator with a computer?
       We would like to avoid 2 FDDI ring and if possible to avoid any.
    3/Is there a better way to design this knowing that we have to separate the
    DATAflows for REAL TIME reasons?
                                                                       

             3 data servers                          12 Workstations
             +-+  +-+  +-+        +---------+        +-+  +-+  +-+
             +++  +++  +++   +----|DECnis   |----+   +++  +++  +++
              |    |    |    |    |Bridge   |    |    |    |    |
             +-----------------+  +---------+   +-----------------+
             |DECconcentrator  |                |DECconcentrator  |
             +--------+--------+                +--------+--------+
                      |                                  |
                      |                                +-+-+ REAL TIME
         Dataflow_1   |                   Dataflow_2   |   | MAIN
       (20 Mbits/sec) |                 (18 Mbits/sec) +-+-+ COMPUTER
                      |                                  |
                      | 6Km = (4 miles)                  | 6km = (4miles)
                      |                                  |
             +--------+--------+               +---------+-------+
             |DECconcentrator  |               |DECconcentrator  |
             ++----+---------+-+   +-------+   ++----+----+----+-+
              |    |         |     +-------+    |    |    |    |
             +-+  +-+        +-----+-------+----+   +++  +++  +++
             +-+  +-+              +-------+        +-+  +-+  +-+
           Other systems    3 dual homing systems   14 REAL time systems
T.RTitleUserPersonal
Name		DateLines
760.1I don't think so, but the design isn't clear to meMUDDY::WATERSTue Oct 27 1992 11:1017
>    1/ Does it work as designed?

    Hard to say.  In the design, which machines are expected to be able
    to talk to each other without resorting to IP routing through REAL TIME
    MAIN COMPUTER?  Next time, label it "COMPUTER/IP ROUTER" if it's meant
    to allow "12 Workstations" and "3 data servers" to talk to
    "14 REAL time systems".

    I gather that the subnet separation points in your design are
    "REAL-TIME MAIN COMPUTER" and "3 dual-homing systems".  I assume
    that "DECnis bridge" is NOT set to filter IP traffic.

    If that's right, the "3 dual-homing systems" won't work for the IP
    protocol, from what little I understand of it.  Dual-homing DAS
    ports should be connected to the same IP subnet, right?  I'm not aware
    that you can dynamically reconnect an IP interface to different subnets
    without the manager's updating the subnet mask of the interface.
760.2More information on the data trafficTOOIS1::MIRGHANEThu Oct 29 1992 08:3359
More information

>>>    Hard to say.  In the design, which machines are expected to be able
>>>    to talk to each other without resorting to IP routing through REAL TIME
>>>    MAIN COMPUTER?  Next time, label it "COMPUTER/IP ROUTER" if it's meant
>>>    to allow "12 Workstations" and "3 data servers" to talk to
>>>    "14 REAL time systems".
	
       
      Here is the traffic:
	- The computers called "Other systems" send messages directly to the 
	3 data servers and only to them.
	- The RT computers send messages to the 12 workstations through 
	the "REAL TIME MAIN COMPUTER" or "IP ROUTER"
	- The 3 computers known as "dual homing"  are in fact connected using 
	2 separated FDDI controllers. (So I think that the term dual homing may
	not be appropriate). On the first FDDI controllers they exchange data 
	with the Workstations through the REAL TIME MAIN COMPUTER.
	On the second FDDI controller they exchange data with the servers.
	- The 12 workstations send data to the 3 data servers too. 
	



>>>    I gather that the subnet separation points in your design are
>>>    "REAL-TIME MAIN COMPUTER" and "3 dual-homing systems".  

	Right:
	- The REAL TIME main computer separate Workstation trafic from 
	REAL TIME trafic. 
        - The "dual homing systems" send 2 types of trafic:
		. to the worksations,
		. to the data servers.

>>>    I assume
>>>    that "DECnis bridge" is NOT set to filter IP traffic.
	The DECnis bridge is there as a data trafic filter:
	As a matter of fact we have:
		-  20 mbits/sec coming from "dual homing" and "other systems"
		   to the data servers.
		- 20 Mbits/sec from the "REAL TIME systems" to the workstations
	The DECnes bridge enables to avoid adding these 2 traffics at the
	workstation level.
 
	 
>>>    If that's right, the "3 dual-homing systems" won't work for the IP
>>>    protocol, from what little I understand of it.  Dual-homing DAS
>>>    ports should be connected to the same IP subnet, right?  I'm not aware
>>>    that you can dynamically reconnect an IP interface to different subnets
>>>    without the manager's updating the subnet mask of the interface.
	In fact, since the so called "dual homing systems" have 2 FDDI
	interfaces each,  each interface is connected to a different IP.
	Will it work then?



Thanks very much for help.
                                                              
Soumetty.