a) It is not possible to use pointers or other types of references in RPCs, as the client and the server may not exist on the same machine and so have different address spaces
b) Data types may have different representations on different systems. For example, long ints are 4 bytes on 32-bit linux systems, and 8 bytes on 64-bit systems.
c) A mechanism needs to be in place for the client stub to determine exactly where the server stub is located (Remote Procedure Call, n.d.).
Why is timestamping needed in Real Time Applications? (This is in context to RTP)
Real time applications use timestamps to reduce jitter. This is especially useful in algorithms using the Leaky Bucket approach, as packets are transmitted at a constant rate; this allows the receiving end to build up a buffer of packets that arrive early, releasing them to the application at that same rate and compensating for subtle increases in latency (Costa, 2008). …show more content…
If the maximum packet lifetime is not large enough to ensure that both the packet itself and its acknowledgements have either been processed or have timed out, the sending process runs the risk of transmitting duplicate packets on congested networks where delivery of a packet or its acknowledgements may be delayed; this means that the maximum packet lifetime must be at least the maximum roundtrip time of a connection, rather than just the maximum time in one direction (Costa,