Our goal for the next iteration is to run a variety of workloads on all databases, and generate graphs to compare each database under different workloads.
We will run each workload with different number of threads, number of nodes, method of data distribution/replication.
If time permits, we will test how well each database fare against failure such as node failure.
Current plan to complete the project (65% done)
- Each group member has their assigned database (Redis, Cassandra, Couchbase, OrientDB, MongoDB) already installed on at least one of five Amazon instances.
- We have also installed YCSB on the instances and ran some initial tests to ensure that inserting, updating, and reading records works for a single instance.
- We are currently working on setting up the distributed configuration for each of our databases.
- As soon as the distributed setup is complete, we expect to start collecting performance data on a small amount of records (e.g. 10,000) and then scaling up incrementally due to limited storage space on the Amazon instances.
- We will then analyze the collected data and compare the performance of each database and represent the differences visually using graphs.
- Time permitting, we will also be investigating how each database behaves/performs in the event of node failure and include our findings in our final report.