Radio resource management (RRM) plays a vital role in wireless networks to efficiently utilize the limited wireless resources and to provide better Quality of Service (QoS). As the radio access networks are continuously enhanced by adding new technology components such as Carrier Aggregation, Flexible spectrum usage, enhanced Inter-cell Interference Coordination (eICIC), and Coordinated multipoint (CoMP) transmission and reception, RRM becomes more sophisticated and challenging. The goal of this project was to design advanced RRM and QoS management solutions that were able to take advantage of new modern technologies and optimize the radio resources to support new services, improve the system capacity, provide better QoS for users, reduce cost and save energy. It included energy-aware resource allocation strategies in LTE uplink, Wireless Network Virtualization, and LTE predictive scheduling approach utilizing fast ray tracing engine.
OpenSAF is a high availability (HA) and management middleware that implements service availability forum (SA Forum) standard specifications. It is an open source software actively supported by leading companies in the communications and enterprise computing industries. OpenSAF has evolved through different versions. Each version of OpenSAF introduces new bug fixes and stability with no changes in the installing and configuration setups. OpenSAF provisions high availability cluster from controller and payload nodes. Controller nodes mainly execute the Availability Management Framework that manages all other payload nodes. Payload nodes are the nodes on which the highly available application software will be running its entities. Installing and configuring the workstations (PCs or VMs) to their roles have been a hassle and time consuming procedure. Moreover, the manual configuration of OpenSAF cluster (HA cluster) hinders the elastic sizing of the HA application in real-time.
Automating installation and deployment OpenSAF is a key enabler for on-demand cluster sizing for HA applications and services. Moreover, it will assure that cluster size is maintained when nodes failed to be recovered from errors. In our project, we used different technologies (MySQL DB, Puppet, JAVA programming, and Ruby scripts) to facilitate the automated deployment of OpenSAF and to assure the cluster resizing without any service interruption. Furthermore, the project will be extended to be integrated with cloud management platforms and HA-aware software management systems to provide a complete solution for elastic deployment of HA-application in cloud. Currently a fully functional prototype have been development and a test bed have been implement to demonstrate an elastic deployment of an IP multimedia subsystem (IMS). The test bed consists of OpenIMS core with modified entities, HA proxy, MySQL DB, Puppet, IMS Bench SIPp, and the tool we have developed.
With the development of the cloud market, cloud computing can be seen as an opportunity for information and communications technology companies to deliver IT services over any fixed or mobile network without violating the SLAs and QoS for end-users. Although cloud computing provides benefits to different sectors in its ecosystem and makes services available anytime, and anywhere, other concerns arise regarding their availability. Unexpected cloud-services outage can have a profound impact on business continuity and IT enterprises. The key to achieve availability requirements is to develop an approach that is immune to failure of applications. Attaining an always-on and always-available service is the main objective of our project. In order to achieve that goal, our work has addressed the problem from different vantage points to generate highly available optimal placement for the requested applications.
We capture high availability (HA) of applications and VMs from different perspectives. Not only the failure rates of applications/VMs affect their HA, but the existence of a backup plan and relations between applications/VMs as well. Besides, the hosting cloud infrastructure has a great impact on HA maximization. Our current research is motivated by the need to design and implement a scheduler that is responsible of resource allocation and applications placement while maximizing their HA. The latter receives end users’ requests and schedules them based on their predefined requirements.
To this end, we developed CHASE, component HA-aware scheduler. CHASE implemented different approaches that deploy redundant models and failover solutions. These practices can be achieved through geographically distributed redundant deployments and considering applications’ intercommunication requirements. However, the tradeoff lies in overcoming the challenges of achieving HA applications/VM placements and compromising between different functionality and failover constraints. CHASE was designed to schedule components in real cloud environment while communicating with OpenStack management system. Furthermore, our work will be extended to have a green-HA scheduler for cloud applications. To this end, we are defining the problem environment, challenges and design methodologies.