Negotiation processes taking place between two or more parties need to agree on a viable execution mechanism. Auctioning protocols have proven useful as electronic negotiation mechanisms in the past. Auctions are a way of determining the price of a resource in a dynamic way. Additionally, auctions have well defined rules such as winner and loser determination, time restrictions or minimum price increment. These restrictions are necessary to ensure fair and transparent resource allocation. However, these rules are limiting flexibility of consumers and providers. In this paper we introduce a novel negotiation based resource allocation mechanisms using the offer-counteroffer negotiation protocol paradigm. This allocation mechanism shows similarities to the supermarket approach as consumer and provider are able to communicate directly. Further, the price is determined in a dynamic way similar to auctioning. We developed a Bazaar-Extension for CloudSim which simulates negotiation processes with different strategies. In this paper we introduce and analyze a specific Genetic Algorithm based negotiation strategy. For the comparison of the efficiency of resource allocations a novel Bazaar-Score was used.

Top