Home

Welcome to the ssd-cache wiki!

ToDo list:

• Generate fake trace with high random data with frequent and sparse time intervals.
• Rerun Contention experiment with different types of Workloads. Can use synthetic workload too.
• Experiment using RI and RD separately
• Get RI more fine grained Tune interval based on RI values. i.e. sudden change in workload should reallocate cache space immediately.
• Think about an experiment for the motivation for multi-tier cache, not contention.
• Look into cplex---optimization software (can we get it to call a function for HR or do we need a mathematical equation?)
• Sim anneal using python libraries. iterative vs. overall
• Scalability of MultiCache. 100's of VMs and 10's of layer. Performance vs. how good?
• Better integration with PARDA

For Tim Stamler:

• Implementation using bcachefs / Xen disk stuff. Senior design project --> Tim (undergrad)
• How to layout cache data on each layer?