Charoy, C. Godart, P. Molli, G. Oster, M. Patten, M. Kuo, A. Fekete, P. Greenfield, J. Web-services Coordination Model; K. Baina, S. Du kanske gillar. Spara som favorit. Skickas inom vardagar. Laddas ned direkt. The Second International Workshop on Cooperative Internet Computing CIC has brought together researchers, academics, and industry practitioners who are involved and interested in the development of advanced and emerging cooperative computing technologies. Cooperative computing is an important computing paradigm to enable different parties to work together towards a pre- defined non-trivial goal.
It encompasses important technological areas like computer supported cooperative work, workflow, computer assisted design and concurrent programming. As technologies continue to advance and evolve, there is an increasing need to research and develop new classes of middlewares and applications to leverage on the combined benefits of Internet and web to provide users and programmers with highly interactive and robust cooperative computing environment.
Each time a request comes, RSU will check whether the dataID is one among the most popular most frequently requested data items, using its request count.
When there is no room to store new data items in cache, a cache replacement strategy should be applied. Cache replacement strategy is to be implemented at two levels; local caches and popular cache. Since popular cache is maintained by considering the popularity of data items, cache replacement is done on the basis of request count by default.
These strategies have very small impact on data availability. The major issue that client cache management faces, is the maintenance of data consistency between the cache client and the data source. All cache consistency algorithms seek to increase the probability of serving cached data items that are identical to those on the server.
However, achieving strong consistency, maintaining the cached items identical to those on the server, requires costly communications with the server to validate renew cached items. Here a weak consistency model based on TTL value of data item is proposed. Each N seconds the entries in the local cache, are invalidated and deleted if they have expired.
Data items in popular cache, maintained by RSU, maintain near strong consistency with the data server.
Each N seconds RSU polls its popular cache to find items in its cache which will be expired within the next N seconds and prepare a prefetch request to the data server to fetch the data item before they get expired. Figure 2 shows the basic interactions between nodes through a scenario in which a vehicle request a data item. First it performs a local cache search.
If it finds such one, DRP packet is forwarded to the corresponding vehicle and data retrieved is send as a DREP packet to the requesting vehicle. Data packets used in the proposed scheme are given in table 1.
Server is the only entity which can update the data, in 2TierCoCS. When a node receives a request for a data item from the end user, it checks whether a valid copy exist in its local cache and returns data to the end user. When a vehicle receives a DREP packet either from RSU or another vehicle, first it checks whether the source is present in the 1-hop neighbour list.
If the source is not a 1-hop neighbor, the vehicle store the data item in its local cache after performing cache replacement if needed, and send CNEW packet to RSU to update Neighbor Cache Index table. In vehicles, as a cache validation mechanism, data items get deleted from local cache, as their TTL get expired. RSU can be viewed as the heart of 2TierCoCs since it participates in most important operations of this caching scheme.
Request count table is used to find the most requested data items and it is updated whenever a new request for data item comes. Each N seconds RSU polls its popular cache to check whether there are items that will be expired in next N seconds and prepare a CURP message which consists of dataIDs of that data items and send to server. Else if it is a SUDP message, the data item itself is updated in the popular cache.
In case of popular cache hit, it returns a DREP packet to the requesting vehicle. If a matching entry is found on Neighbour Cache Index, DRP packet is forwarded to the corresponding vehicles carrying data. Another procedure to be accomplished, when a DRP packet is received, is to update request count table and sort the table according to request count. If the newly requested item is one of the maximum items, then a DRP packet is forwarded to the server to fetch data and place in popular cache. Else the DREP packet is forwarded to the destination vehicle.
A theoretical analysis is done here to study the effect of procedures performed on each entities of proposed scheme.
ICOMP (Jul ), International Conference on Internet Computing, Las Vegas USA - Conference
Run-time complexity of each procedure is theoretically analyzed. Most of the procedures are accompanied with a search algorithm and takes O P time in worst case, where P is the size of list to be searched. Cache discovery takes O K time in its worst case, that is when data is found on neighboring vehicles, where K is the number of entries in Neighbour Cache Index.
Local cache of size N takes O N time to announce a hit or miss. Popular cache takes O M time to search the required data in M items. Neighbour Cache Index search takes O K. Cache replacement takes O N time in its worst case, since it include searching the list of size N, for the data item with least TTL.
- The 2nd International Conference on Computer Communication and the Internet (ICCCI )!
- Hours of Operation.
- Contact Information!
- Gcc notes unit 1.
In Cache admission procedure, source address is searched in the neighbor list of H items, and takes O H time. In vehicles cache consistency check is of time complexity O 1 , since each data item is deleted when the TTL is expired. When considering complexities of each entity, server, vehicle and RSU, each of them comprises a search operation and time complexity is first order. VANETs are emerging network technology, whose main purpose is to ensure traffic safety. So when handling internet data, VANETs face problems in data availability due to frequent disconnections and large network traffic.
ICOMP'10 - The 2010 International Conference on Internet Computing
Cooperative caching is an appropriate solution. Here a two-tier cooperative caching scheme for VANETs 2TierCoCS is proposed, which address the issues of cache discovery, cache admission control, and cache consistency. In the proposed scheme caching is done at two levels of architecture; vehicles and RSU, so that hit ratio can be improved. The cooperative caching scheme proposed here, will improve the data availability in VANETs, by reducing query latency and bandwidth usage, by avoiding the need for data from the Internet data server.
For future work, 3 directions to extend 2TierCoCS can be explored. In current scenario, cached copies of data items in the proposed cooperative caching scheme have a weak consistency, which is maintained by assigning a TTL value. Consistency level can be enhanced by using better models of consistency, even without introducing large traffic on the network.
Another challenge faced by this scheme, arises when the vehicle, that requested data, leaves the zone of RSU, before the DREP packet arrives. Routing of data packets may need handover among RSUs. A new architecture similar to cellular network can be developed, which addresses aforesaid issue of packet delivery. Since the proposed scheme is a distributed scenario, to analyze the real effect of these distributed algorithms, a probability model has to be applied. So finding and fitting an appropriate model of probability can be done to enhance the analysis of this scheme.
Implementation of this scenario in real environment can be done as a future work.
An evaluation of vehicular networks with real vehicular GPS traces. Prabhakar Ranjan, Kamal Kant Ahirwar. Preetha Theresa Joy,K. Poulose Jacob. IEEE ,, Liangzhong Yin and Guohong Cao. Yu Du and Sandeep K.
Therenece Houngbadji and Samuel Pierre. Lilly Sheeba S and Yogesh P. Mieso K.