[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

23.0. "Connect other topologies?" by DENVER::CASE (SWS DNT DVO) Tue Jan 02 1990 13:36

    I'm trying to research the potential problems with linking other
    topologies, particularly 802.3 and 802.5, to FDDI, and thru FDDI, to
    each other.
    
    It seems to me that FDDI will provide an excellent backbone and central
    transport to connect many LAN topologies, not just 802.3.  And since
    FDDI handles larger packets than either 802.3 or .5, it should be
    relatively easy to bridge 802.3 to FDDI to 802.3, or 802.5 to FDDI to
    802.5.
    
    But once other vendors begin marketing 802.5-FDDI bridges, and both
    802.3 and .5 are linked to the same FDDI ring, customers will naturally
    want to interconnect stations on each topology to the other.
    
    One concern, then, is that if a 4KB 802.5 packet gets routed thru FDDI
    to an 802.3 bridge, is there any way to break up the packet into
    multiple, smaller 802.3 packets?  Will our bridge have this capability?
    
    Addressing and frame headers could be handled via 802.1 and 802.2 LLC. 
    I've heard concerns in other 802.3-.5 linkages about bit-ordering
    within the data portion of the packet ("Big-endian" vs
    "Little-endian").  Is it possible to, and have we considered, handling
    this problem in the bridge?  Or is it a higher-level protocol issue?
    
    Any concerns/difficulties with addressing across multiple bridges? 
    Will interconnection via FDDI be limited to common protocols (e.g.
    TCP/IP, OSI)?
    
    Any discussion and ideas in this area would be appreciated.
T.RTitleUserPersonal
Name		DateLines
23.1What KINDS of bridges?AKO569::JOYSo many men, so few with brainsTue Jan 02 1990 14:0412
    One of the biggest issues with bridging 802.* protocols to FDDI will be
    if the bridge is designed to be a translating bridge or an
    encapsulating bridge. Fibronics bridges of today are encapsulating, so
    they aren't nearly as versatile as our translating bridge will be as
    far as connectivity between nodes on the FDDI ring or the Ethernet LAN.
    I would imagine that if all translating bridges are used on the FDDI
    ring, then there shouldn't be a problem of connectivity/message-passing
    between various other LAN technologies, i.e. 802.3 to 802.5 with the
    FDDI ring as the go-between. 
    
    Debbie
23.2Applies to Data?MORK::CASESWS DNT DVOTue Jan 02 1990 16:5012
    Good point.  I was aware that our bridge will be translating, but does
    this mean that it could handle the encoding or ordering of the data
    portion of the packet?  I though that the translation function would
    handle the addressing and packet header(s), translating from one
    standard to the other, but ignore the data portion of the packet and
    pass it intact to the destination address.  If so, we would have a
    problem with the destination thinking it's reading 802.3-style data,
    when it's really from an 802.5 source.
    
    I'll do some more checking in other notes files on bridges in general,
    as I'm sure the issues could pertain to bridging any dissimilar
    topologies.
23.3Only at the ESKYOA::MENNERPeople everywhere just wanna be freeFri Jan 05 1990 01:454
    
    The correct presentation of user data involves a higher level protocol 
    (for example: OSI layers 6 & 7) which would be implemented at the end 
    systems only (and not at bridges/routers).
23.4Where do the bits go?SUBWAY::BRIGGSHave datascope, will travel.Mon Jan 08 1990 17:0810
    
    If a directly attached FDDI system sends a packet larger than
    max Ethernet frame size to an Ethernet segment is the packet
    truncated or somehow fragmented.
    
    Stated differently, what is the translating 10/100 bridges policy
    in forward this sort of packet?
    
    I realize that this is not a major concern because higher lever
    protocols should set the maximum {packet, segment, ru...} size.
23.5KONING::KONINGNI1D @FN42eqMon Jan 08 1990 19:0314
    802.5 has a larger maximum message size than FDDI; only Ethernet has a
    smaller limit.
    
    As for bit order, the real issue there is between FDDI and Ethernet,
    since FDDI follows the 802.5 backwards bit order.  The answer is that
    the bridge will do the right thing here, so that both addresses and
    data come across in the way the various networks expect to see them.
    (There's a caveat: if addresses are contained in the data portion of a
    message, then the bridge doesn't know about that, of course.  For those
    cases, it's up to the various protocols to agree on a standard
    encoding.  For the most important ones, such as ARP, that agreement has
    been reached.)
    
    	paul