Resource Allocation Tools

Resource Allocation Tools


Yan LiPKUliyan05 at
Guo YangPKUyangguo10 at
Fenghong ChenPKUchfh10 at

Introduction to Project

In a shared cluster, each application runs on a subset of nodes and these subsets can overlap with one another. Resource management in such a cluster should adaptively change the application placement and workload assignment to satisfy the dynamic applications workloads and optimize the resource usage. This becomes a challenging problem with the cluster scale and application amount growing large. To solve the problem, we designed a tool to manage the resource in a JOn2AS cluster, and build a platform to support the applications running in the cluster. In the tool, we 

  • Propose a novel self-adaptive resource management approach with quick responsiveness, high scalability, and application prioritization in addition to managing the resources effectively. 
  • Implement OSGi-based middleware framework with dynamic configuration of services. The parameters, types and specific implementation of middleware services can be configured at runtime according to the application requirements, thus avoid unnecessary waste of resources; 
  • Implements a proposed mechanism for resource partitioning with the virtual machine technology, making the middleware can not only control the node resources, but also isolate nodes' runtime environment. 

The critical technical point of the resource management tool is as follows: 

  1. Based on the variation forecast of requests and the free resource situation, give a plan to allocate the resource dynamically. 
  2. According to the results of the allocation of resources planning, allocate the resource in the cluster, dispatch the requests to each node. 
  3. Configure the application runtime environment dynamically depends on the application demand. 
  4. Provide a management console, to observe the operational status of the shared cluster and resource allocation.