TABLE OF CONTENTS This section gives a brief discussion of possible, future, overload controls.
SIP [26] and HTTP [27], from which it is derived, do not currently have any overload control methods that would meet the requirements presented here. However, they both have elements that could support the incorporation of overload controls designed according to section 7.6.
SIP is an application-layer control protocol that can establish, modify and terminate multimedia sessions (conferences) or Internet telephony calls. In particular, after sending an INVITE request message from a client to request participation in a session, there are various response messages returned which can be used in feedback control. Some of these relate to the progress of a request due to a server, and so can be used for server (nodal) load control. Others relate to a destination server status, e.g. busy or unavailable, and so can be used for overload focused on to a specific destination (using as identifiers To Header fields or Request-URI ). Staged overload control can be applied because so-called intermediate nodes or proxy servers are used, for example, to route requests, enforce policies, and control firewalls.
HTTP has a similar structure, and therefore similar techniques can be used for restricting the rate of requests for a Web page. However, in the context of the WWW, caching [28] is an additional mechanism available to control overload to static pages. Rather than just restricting the rate of requests, copies of pages can be distributed away from...
