CloneCloud: Elastic Execution between Mobile Device and Cloud

The demand for mobile devices is continuously on the rise. People want their mobile phones to do the same tasks as their laptops and desktops. But the nature of the mobile phones hardware still limits them when it comes to sheer processing power. To tackle these issues the paper proposes a system called ‘Clone Cloud’ which as the name suggests creates a clone of the mobile device on the cloud, which can be used to offload heavy computational tasks.

Main Points:

• The model is primarily applicable to application layer virtual machines, since the instruction sets in it are byte codes it is hardware and OS independent. This makes the cloud cloning smooth and fast.
• The partitioning component for finding migration points uses – static analysis to find the constraints and the dynamic profiling for building the cost model for execution and migration. It also uses an optimizer that uses the above constraints and cost models to derive the partitions.
• CloneCloud uses an off-line partitioning mechanism to divide parts of an application for local and remote execution and stores the corresponding results in a database.
• The static analyzer identifies the legal partitions of the application executable according to a set of constraints Migration is restricted to the method entry and exit points.Migration is allowed only at the boundaries of application methods but not core system library method. It’s also allowed at the VM-layer method boundaries but not native method boundaries.
• The Dynamic profiler collects data that will be used to construct the cost model.Each execution is run once on mobile device and once on the clone in the cloud.The profiler outputs set of executions S and a “profile tree”, for both mobile device and the clone to calculate the cost.
• The two main features of CloneCloud for distributed execution are thread granularity migration and native-everywhere operation which are very useful in specific applications.

Trade-offs:

• The paper does not comment about any security issues like data security during the cloning activity.
• The offloading mechanism is significantly different from MAUI.
• Even though the paper proposes thread granularity it does not support multithreading.
• The results given in the paper are based on testing the CloneCloud system on a single mobile device. I assume results may vary if the system is tested on multiple devices.
• The MAUI paper focuses a lot on RTT in the result section in comparison the CloneCloud paper has no significant mention of RTT.