

<?xml version="1.0" encoding="UTF-8"?>
<record>
  <title>A Concept Recommender Service on The Federated Cloud of the PPP European Project</title>
  <journal>Journal of Intelligent Computing</journal>
  <author>Boris Moltchanov, Oscar Rodriguez Rocha</author>
  <volume>6</volume>
  <issue>1</issue>
  <year>2015</year>
  <doi></doi>
  <url></url>
  <abstract>Conventionally, the Web Service [1], respecting the Client-Server paradigm, consists of a software system or a
platform running as a server somewhere in the Internet allowing its customers to benefit from it through the use of a client.
Cloud computing [2] (which has gained a lot of popularity in recent years), allows to have this software system or platform
deployed and running in a flexible and robust environment where traffic demand is automatically handled and the faulttolerance
and high-availability are provided by the cloud tools and means. However, currently existing private and public
cloud models comprise cloud platforms installed on IT facilities located in certain geographical areas. Thus, the final customer
location is not considered and the service traffic is not balanced geographically. This is a serious drawback of the existing
technology as it creates an overhead of the network (Internet) traffic for the service when the customers use it from different
geographical areas. Moreover this increases service delays as the packets in the request-response communication paradigm
cross long Internet cross-countries connections and thus delayed. The new concept proposed in this paper is feasible due to the
EU attempt to instantiate a number of the public clouds interconnected into a cloud federation over Europe, where the
deploying of services in a geographically distributed cloud environment will enable to serve customers in a more efficient and
faster way. The location of the customer connected through the network (Internet) will define to which service instance in the
federated cloud, its requests will be sent to. This drastically reduces service traffic, and so the costs, over Internet for international
traveling customers, makes the service reaction faster as it balances the service traffic over mode nodes and makes
significantly more robust overall service solution because of the service geographical redundancy. The recommender service
described in this paper is deployed over few nodes of the federated clouds build and interconnected as result of the EU funded
PPP XIFI project[3].</abstract>
</record>
