Benefits and limitations of tapping into stored energy for datacenters
Proceedings of the 38th annual international symposium on Computer architecture
Leveraging stored energy for handling power emergencies in aggressively provisioned datacenters
ASPLOS XVII Proceedings of the seventeenth international conference on Architectural Support for Programming Languages and Operating Systems
Aggressive Datacenter Power Provisioning with Batteries
ACM Transactions on Computer Systems (TOCS)
| Hi-index | 0.00 |
We introduce a dynamic battery model that describes the variations of the capacity of a battery under time varying discharge current. This model supports a co-design approach for cyber-physical systems where control laws,online scheduling algorithms and battery management can be designed jointly to balance performance of the plant,the computing device, and the batteries. The input-output relationship of our model is equivalent to the Rakhmatov-Vrudhula-Wallach (RVW) model that has been verified by high fidelity simulations and experiments. The major improvement of our model is that it allows online prediction of battery life. This improvement is significant since for cyber-physical systems, control laws and online scheduling algorithms draw current from batteries according to the state of the plant and the processor, which cannot be determined at the design stage. Using our model, the optimal discharge profile for a square wave current are determined and simulation results are provided to compare battery life for different discharge profiles.