Most Social networks at the moment provide information through APIs using json, jsonp or XML serialization. This reaches the third out of five stars on the linked data star rating system . In our architecture we wanted to provide four and five star on linked data rating system by publishing the users’ data in RDF format. In this way people who give permission to do so they are able to point to movies they watched, liked or commented. In that way software agents can take advantage of this information and make suggestions like, movies that me and my friend does not have seen and combined with our agenda information the agent is able to suggest cinemas and hours and movies that we might like.
There are different ways to publish RDF data. The first is called Native stores. The information in these stores is stored directly in to the hard drive implementing in this way a RDF(S)/OWL oriented database. Some examples of Native stores are Kowari  and OWLIM  . The second approach is called DBMS-Backed. These systems employ mostly relation databases like MySQL, PostgreSQL etc. to store RDF triples. Some examples of these systems are YARS  and RDF Suite  . The third approach is the RDF wrappers where software components are stored above databases with ordinary relational information and transform relational data into RDF triples. Some examples of RDF wrappers components are D2R Server  and Triplify . Finally the forth approach is the Hybrid stores where a modular architecture is provided enabling the use of different store types. Some examples of hybrid stores are Jena  , Sesame , Virtuoso and Redland .
In general, stores that use extensively indexes (Yars,Virtuoso, Kowary) result in faster query answering but they consume more memory . D2RServer and Triplify RDF wrappers lack on SPARQL facilities. More specifically D2RServer does not provide SPARQL endpoint at all where Triplify the query response time is on averate slow . Finally in-memory storage provided by Jena and Sesame have better performance in comparison with stores that store data on the hard drive. The downside of it is that the amount of data cannot exeed the size of the computers main memory  .
In our case we need the relational database structure due to the fact that at the moment relational database have better performance in comarison with the Semantic Stores. For this reason the best option is to use RDF wrappers using the best of both worlds.
3 Wood, David, Gearon, Paul, and Adams, Tom. Kowari: A platform for semantic web storage and analysis. In 14th International WWW Conference (Chiba 2005).
4 Kiryakov, Atanas, Ognyanov, Damyan, and Manov, Dimitar. OWLIM – A Pragmatic Semantic Repository for OWL. In 6th International Conference on Web Information Systems Engineering.
5 Harth, Andreas and Decker, Stefan. Yet Another RDF Store: Perfect Index Structures for Storing Semantic Web Data With Contexts. DERI, 2004.
6 Alexaki, Sofia, Crisophides, Vassilis, Karvounarakis, Greg, and Plexousakis, Dimitris. On Storing Voluminous RDF Descriptions: The case of Web Portal Catalogs. ( 2001), ACM.
7 Bizer, Christian and Cyganiak, Richard. D2R Server – Publishing Relational Databases on the Semantic Web. In 5th International Semantic Web Conference ( 2006).
8 Auer, Sören, Dietzold, Sebastian, Lehmann, Jens, Hellmann, Sebastian, and Aumueller, David. Triplify: light-weight linked data publication from relational databases. In 18th international conference on World Wide Web (Madrid 2009), ACM, 621–630.
9 McBride, Brian. Jena: A Semantic Web Toolkit. IEEE Internet Computing, 6 (2002), 55-59.
10 Broekstra, Jeen, Kampman, Arjohn, and Harmelen, Frank van. Sesame: A Generic Architecture for Storing and Querying RDF and RDF Schema. In International Semantic Web Conference (2002), 54–68.
11 Erling, Orri and Mikhailov, Ivan. RDF Support in the Virtuoso DBMS. In 1st Conference on Social Semantic Web (CSSW) ( 2009), Springer Berlin / Heidelberg, 7-24.
12 Beckett, David. The design and implementation of the redland RDF application framework. In 10th international conference on World Wide Web (Hong Kong 2001), ACM, 449–456.
13 Liu, Baolin and Hu, Bo. An Evaluation of RDF Storage Systems for Large Data Applications. Semantics, Knowledge and Grid, International Conference, 0 (2005), 59.
14 Minack, Enrico, Siberski, Wolf, and Nejdl, Wolfgang. Benchmarking Fulltext Search Performance of RDF Stores. In Aroyo, Lora et al., eds., The Semantic Web: Research and Applications. Springer Berlin / Heidelberg, 2009.
15 Ma, Li, Yang, Yang, Qiu, Zhaoming, Xie, Guotong, Pan, Yue, and Liu, Shengping. Towards a Complete OWL Ontology Benchmark. In 3rd European SemanticWeb Conference (Budva 2006), Springer Berlin / Heidelberg, 125-139.
16 Bizer, Christian and Schultz, Andreas. Benchmarking the performance of storage systems that expose SPARQL endpoints. In Workshop on Scalable Semantic Web Knowledge Base Systems (Karlsruhe 2008)
17 Smith, Daniel Alexander, Owens, Alisdair, schraefel, mc, and Sinclair, Patrick. Challenges in Supporting Faceted Semantic Browsing of Multimedia Collections. (Genoa 2007), Springer-Verlag New York Inc.
18 Janik, Maciej and Kochut, Krys. BRAHMS: A WorkBench RDF Store and High Performance Memory System for Semantic Association Discovery. ( 2005), Springer Berlin / Heidelberg, 431–445.
19 Guo, , Yuanbo, Pan, Zhengxiang, and Heflin, Jeff. An Evaluation of Knowledge Base Systems for Large OWL Datasets. In 3rd International Semantic Web Conference (Hiroshima 2004), Springer Berlin / Heidelberg, 274-288.