Discuss whether it is possible to improve upon the ‘pull’ model of service discovery by multicasting (or broadcasting) and caching replies to queries.

What will be an ideal response?

We start with the ‘pull’ model described in the chapter but turn it into a hybrid with the ‘push’ model by having services multicast/broadcast their existence when a query occurs. All potential client devices can then hear and cache the service advertisements. When a component on the device requires a service, it first consults the local cache. If no matching services are found or if a more up-to-date list of matching services is required, then the device still multicasts its query. However, now it includes information about any matching services in its cache. Other potential clients may now learn about those services’ existence. Moreover, only services not in the query need reply.

While it is intuitively reasonable, this hybrid scheme needs validation through simultaion or experimentation against the required implementation parameters. The extra multicasts incur a cost.