Service Placement in Ad Hoc Networks by Georg Wittenburg & Jochen Schiller

Author:Georg Wittenburg & Jochen Schiller
Language: eng
Format: epub
Publisher: Springer London, London

The inquiry begins with a REPLICATION REQUEST packet, which contains information regarding the resource requirements of the service. Upon reception of this packet, the target node ensures that the service in question is not already running on this node and that enough local resources (e.g., memory, processing capacity, etc.) are available. If one of the conditions is not fulfilled, the target node replies with a REPLICATION NACK packet. Depending on the placement algorithm employed on the current service host, it may then take note of the refusal of this target node, e.g., by removing it from the set of potential replication targets for future placement decisions (cf. Sect. 3.3.3.1).

If, however, both conditions are fulfilled, the target node accepts the replication, thereby moving on to the second phase of the replication protocol. Based on the information embedded in the REPLICATION REQUEST , the target node initiates the transfer of the service data by sending a REPLICATION ACK packet to the current service host. A REPLICATION ACK contains fields specifying which segment of the service data the target node expects the current service host to transmit. This may either be one or multiple segments, thereby allowing for flow control between target and originator node. If the current service host receives the REPLICATION ACK packets, it sends one or multiple REPLICATION DATA packets that contain the requested service data. Once the complete service data (as specified in the initial REPLICATION REQUEST ) has been transferred to the target node, it sends a final REPLICATION ACK requesting no further data to the originator node. This indicates to the current service host that the service replication was successful, and that it may proceed to resume normal operation or, in case of an ongoing service adaptation, move on to implement the next action of that adaptation. Finally, the target node proceeds to start its own service instance.

This protocol for service replication leaves the control of the process on the receiving side at the new service host. Once it has agreed to host a new service instance, the target node can thus easily implement flow control and request retransmissions of lost packets. The current service host does not need to hold any state regarding this process, except for a timer to abort and unlock the service instance in case of the target node becoming unavailable before the replication is completed.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.