What the heck is a Cyber-Physical System?
Wikipedia says:
Unlike more traditional embedded systems, a full-fledged CPS is typically designed as a network of interacting elements with physical input and output instead of as standalone devices.
Examples of CPS include smart grid, autonomous automobile systems, medical monitoring, industrial control systems, robotics systems, recycling and automatic pilot avionics.
So basicly, CPS is a fancier term for IoT. The emphasis is on "cyber", i.e. a netwrok of devices as a whole.
What is the key problem this paper trying to address?
How to achieve fault-tolerant and high-performance in the mean time in a storage system.
Comments
Summary
This paper tries to address the fault-tolerance and performance issue in a distributed storage system. Fault-tolerance is achieved through replicating data items to multiple nodes, with each node given by consistent hashing. High-performance is achieved by saving disk IO in lookups via a in-memory bloom-filter.
However, it's unclear what are the differences in the problem considered or the technique used compared to the state-of-the-art. This paper also does not include any experimental evaluation comparing the state-of-the-art.
Comments
- It's unclear what are the main contributions.
- It's unclear what is different between CPS storage considered by the authors and well-studied distributed storage systems.
- It's unclear what is different between the proposed consistent hashing and those in the literature, such as the one in the dynamo paper.
- This paper does not show any performance comparison with the state-of-the-art.