could be the small engine (s), type: scale online social networks

Google Tech Talk (see below) June 17, 2010 Presented by Josep M. Pujol. SUMMARY The social graph partitioning problem has new challenges of the planning system for the scale of online social networks (OSN). vertical scale for the use of full replication can be a costly affair. The scale can horizontally by partitioning and distributing data across multiple servers, using as key-value is stored using DHT suffer from communications between servers and expensive to provide differentArguments. Such challenges have often resulted in costly efforts to re-architecture for OSN popular as Twitter and Facebook. We design, implement and evaluate SPAR, partitioning, online replication and middleware that mediates between application and database layer of OSN. SPAR uses the underlying structure to guarantee their social graph to partition data and selectively replicate user that the user their neighbors, co-located data on your computer. The profits of these are multi-fold:Application developers can expect local semantics, namely the development as a single machine, scalability is achieved with the addition of commodity machines with low memory requirements and network I / O, and N + K redundancy becomes reached a fraction of the cost. We offer a complete design of the system, overall evaluation of records from Twitter, Facebook and Orkut, and an implementation of the work. We show that SPAR also leads in terms of reducing overhead costs and dealing with high dynamic...