imp@jhereg.osa.com (Charles T. Lukaszewski) (10/25/90)
In article <1990Oct18.172918.24929@forwiss.uni-passau.de> grass@ unipas.fmi.uni-passau.de (Grass) writes: > >This sounds that at least group address recognition is done by software: all >messages with its first destination address bit set are received by every >station, transferred to some place in memory and examined by some software >routines. If the destination address does not match with the actual set of >active group addresses the whole message is discarded from memory. What a >waste of cpu/memory/bus ressources! One implementation that has not been mentioned in this discussion is the Digital VAX Ethernet controller and the DECnet Phase IV protocols. Though reliable information is hard to come by, I have studied the Phase IV proto- cols at length with HP network analyzers. It turns out that a VAX, especially a member of a Local-Area VAX Cluster, actually uses a series of Ethernet addresses, with the DECnet software "reprogramming" the board's MAC address regularly. DEC makes extensive use of multicasts, and you can distinguish between various types of con- versations based on the first octet in the destination address (AA and AB are the primary types). In fact, Ethernet addresses beginning with AA seem to be addressed to particular stations, as the DECnet node number and area number are also encoded into the address. Hewlett-Packard has a really excellent performance analysis package for DECnet which offers the same kind of statistics as are available for general Ethernet traffic - FYI. I'd like to hear from someone more knowledgeable about DECnet internals - reverse engineering is such a frustrating excercise. -- _______________________________________________________________________________ Charles T. Lukaszewski imp@osa.com 612 525-0000 Managing Partner & Chairman Open Systems Architects, Inc. "Who needs a disclaimer? I liked the opinions so much, I bought the company!"
oberman@rogue.llnl.gov (10/25/90)
In article <1990Oct25.004509.7608@jhereg.osa.com>, imp@jhereg.osa.com (Charles T. Lukaszewski) writes: > > One implementation that has not been mentioned in this discussion is the > Digital VAX Ethernet controller and the DECnet Phase IV protocols. Though > reliable information is hard to come by, I have studied the Phase IV proto- > cols at length with HP network analyzers. > > It turns out that a VAX, especially a member of a Local-Area VAX Cluster, > actually uses a series of Ethernet addresses, with the DECnet software > "reprogramming" the board's MAC address regularly. DEC makes extensive > use of multicasts, and you can distinguish between various types of con- > versations based on the first octet in the destination address (AA and AB > are the primary types). In fact, Ethernet addresses beginning with AA > seem to be addressed to particular stations, as the DECnet node number > and area number are also encoded into the address. > > Hewlett-Packard has a really excellent performance analysis package for > DECnet which offers the same kind of statistics as are available for > general Ethernet traffic - FYI. > > I'd like to hear from someone more knowledgeable about DECnet internals - > reverse engineering is such a frustrating excercise. The DECnet internals are fully documented and available (for about $500) from DEC. DECnet does not reprogram the MAC address "regularly". It does it only once when DECnet is started. It will do it at boot time if SCSSYSTEMID is non-zero. The physical address is reset to AA-00-04-00-nn-nn where nn-nn is calculated from either the SCSSYSTEMID or the DECnet address. (If SCSSYSTEMID is non-zero, it MUST match the DECnet address in the form (1024*area)+node.) This is from the old idea used in both DECnet and XNS that a system should have the same address for all ports. This idea isn't work out real well and no other protocols that I am aware of do this. DECnet does make heavy use of multicasts, but a first octet of AA is not a multicast. AB is. The group address bit is the least significant bit of the first byte, si that if the first byte is odd, it is a multicast. As the signal comes over the wire, this is the first bit seen by the interface. DEC does not EVER use broadcasts for their own protocols. They consider broadcasts to be evil. (So do I.) Most protocols now use multicasts and not broadcasts. ARP is an exception. While it could use multicasts perfectly well, it will take a lot of effort to get all of the systems converted at this late date. R. Kevin Oberman Lawrence Livermore National Laboratory Internet: oberman@icdc.llnl.gov (415) 422-6955 Disclaimer: Don't take this too seriously. I just like to improve my typing and probably don't really know anything useful about anything.