The IP over ATM Mailing List Archive by date[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] If broadcast is a special case of multicast, then....
Curtis, > I think it will be impossible to force that ALL applications > can not use broadcast into LIS. > Once some application has to use broadcast, there may be three solutions. > > i) we do not support such a application cv> It is not inconcievable to have a host that does not support a cv> particular application that relies on broadcast (netBIOS or Lotus cv> Notes come to mind or even unmodified bootp) yet sits on a LIS that cv> does. cv> Ideally any protocol would use a multicast to reach only those hosts cv> that are running the protocol. There may be a set of protocols that cv> continue to use broadcast. It may be common, particularly when all cv> but the most baroque have been converted, for hosts not to participate cv> in any of the protocols still using broadcast. By my stupid understanding, there are two way to multicast the message to the entities who use the same protocol, WITHOUT prior address knowledge in the client application. The network has infomation of members associated with every protocol to establish every multicast connection. This information will be (1) registered by the client or be (2) statistically configureed by the system operator. In order to use (1), we must have a well-known registration server(s) address, and redistration server must maintion address information for EVERY protocol. I should say this approach will be possibly work, while having no network/host specific address information. But, having a broadcast channel (this could be small bandwidth and could be unnecessary to provide a so reliable data delivery service) will be much easier, if the provision of boradcast service for LIS is not hard to provide. Many cleaver applications/host will not frequently use multicast/broadcast service. Then, why the provision of broadcast will be hard compared to the provision of muticast service ? Regards, Hiroshi Esaki |
|