A Quorum Based Content Delivery Architecture.
Michael Kapralos, John A. Chandy · 2005
In this ever-growing world, the amount of information obtained by the human race is increasing at an exponential rate. However, along with this progress comes the need to develop the infrastructure to make this information accessible. Since a vast number of people are interested in accessing this information from a relatively small number of servers, an efficient algorithm must be developed and implemented to resolve this dilemma. It would be ideal that every user would be able to access only one server to get all the information. However, this is impractical due to storage limitations. It would also be inefficient because the data would have to be replicated and stored throughout multiple servers whenever a change was made to any or all of this information. This also leads the problem of mutual exclusion. If a server was updating the information, no information would be available to any other client on any other server. To make this information readily available, a new quorum-based scheme for managing this replicated data is developed and described in this paper. This quorum-based scheme will divide the servers into groups containing redundant data, and ensuring the clients that they can access information quickly without compromising network quality, also limiting the number of servers which must be read as well as written.