A User Experience-Based Process Resources Sharing Management Mechanism In The Cloud

As a new type of resource usage mode and delivery mode, cloud computing has got more and more attention by all walks of life. It has become a popular technical term now.Cloud computing solves the wastes of the resources caused by the formerly exclusive use of computer resources. The huge resources pool in the cloud offers the possibility for users to choose and share various resources. At the same time as its technology is maturing, data share and application share can be easily achieved between different devices, hence making it more convenient to use cloud resources.Business process allows a variety of loosely-coupled services combined to capture business logic in an abstract and reusable way, and thereby BPM plays an important role in many organizations. With the features of economy and convenience, cloud computing attracts increasing numbers of manufacturers to utilize the cloud resources to manage their business processes.The modern life puts forward the requirement of more quickly and higher resources utilization of cloud computing. However, the frequent used mode of process engine sharing is not flexible enough. We want to find out a new mode to share the resources to their greatest degree and make the management more flexible. Among the resources sharing model, the one of Single-Business-Process-Instance Multi-Tenants is the way to share the resources to their greatest degree, hence maximizing their economic benefits. For this reason, we aim to find a way to use the Single-Business-Process-Instance Multi-Tenants in the cloud environment. Moreover, the higher the sharing scale of the process nodes, the heavier the load they bore, we aim to find a new way to solve the balance problem in process sharing in the cloud.Based on the requirement we mentioned above, our basis new ideas are as follows:1. We put forward a multi-tenant process resource sharing management framework, and give a detailed introduction of the resource management process both on runtime and in the deployment phrase. Our design avoids the redeployment of the whole process, hence give convenient to the use of the process.2. We study depth in the balancing problem, analyzing the cause and the solution. We present a load balance strategy that dual optimize users’experience from both deployment and runtime. This load balance strategy is composed by dynamic replica placement strategy and optimal process node selection strategy.3. We design and implement a Pub/Sub-Based distributed resources management platform. Based on this platform, the way to decouple the whole process is introduced. The multi-tenant message routing mechanism of the Pub/Sub network is laid out. We also demonstrate the flexibility and the performance of the process resources management pattern from several aspects, such as response time, throughput, system usage, and response lost.The Pub/Sub-based multi-tenant process resources management framework we proposed could maximum the resources sharing as well as solving the balance problem for every tenant. Based on this framework, a load balance strategy that dual optimize users" experience from both deployment and runtime can manage the process effectively.