摘要:
As companies are moving towards cloud infrastructure and big data, a tradition- al single database server cannotsatisfy their needs. NoSQL databases are considered as an alternative solution to this situation. NoSQL provides high performance, elas- tic scaling and low cost for storing large amount of data. However, there are many NoSQL databases to choose from. Each has its advantage and disadvantage. In this thesis, we evaluate several popular NoSQL databases including: Redis 1, HBase 2, MongoDB 3, and Cassandra 4 on their performance against several workloads. In addition, because the current version of Redis does not provide scalability the same way as HBase or Cassandra does, we introduce Dynamic Redis that extends Redis and allows the dynamic addition and removal of Redis servers without impact- ing the running system. Our implementation also takes the advanced interface of Redis that includes list, and set management into account.