The Copperbelt University Computer Centre runs five physical servers 24/7 throughout the academic year calendar. These machines consumed a lot of resources such as electricity used to run and cool them. In addition, the Computer Centre employed a lot of technicians to run, maintain and service the named servers. All the discussed costs were incurred throughout the year including the idle workload period when there was very little work to be processed. It was against that background the Computer Centres Resource Cloud Elasticity-Scalability (CRECES) was envisaged to only use one physical server during idle workload period. It was at such periods when the Computer Centre carried a rapid elasticity to scale down on the huge server resource utilisation. Similarly, at peak workload time the centre carried a rapid elasticity to scale up by producing many virtual hardware and software resources just from a single server to five virtualised ones. Hence, the capacity of the Centre to scale up and down resources acted as a cost serving measures in utilising the hardware and software resources.
TopIntroduction
The larger capital expenditure of the Copperbelt University comes from the running of the Computer Centre, which provides the information communication technology (ICT) or information technology (IT) services to the entire institution community. Currently the Centre has five (5) physical servers that have been consuming a lot of expenditure on electricity which is used to power on the machines 24/7. The electricity is also used to cool the five servers and room ventilation. In addition, the Computer Centre employed a lot of technicians to run, maintain and service the named servers. All the discussed costs are incurred throughout a twelve month academic year calendar including the idle workload period. All this causes overheard costs even when the same number of machines is running whilst there is less workload being processed. It is in view of that the Computer Centres Resource Cloud Elasticity-Scalability was envisaged and in this work referred to and abbreviated as CRECES. The CRECES scales up or down the hardware and software resources either during the peak or idle workload periods to reduce the cost of production. For instance during the idle period that stretches from April to October, the Computer Centre shall only use one server in this case referred to as host server. In this way, fewer resources shall be used, hence very little electricity will be required to run a single machine. During the peak workload period, the host server shall be virtualised, to yield five guest servers that will process all the institution workloads at a low cost of production.
The CRECES has an ‘inthahs’ which is composed of the Host Operating System, Type 2 Hypervisor, and Guest Operating System. These technologies during the peak period produce five virtual guest servers that process huge amount of workload. These guest servers operate imitating the host machine. At this time, all the university workload is simultaneously processed. The other benefit of the CRECES, is that less manpower shall be maintained and this shall reduce the labour costs. As currently, the university still pays same salaries to the large workforce even during the idle workload period.
Problem Statement
The Coperbelt University like any other academic institutions in the sub-Saharan region is none profit making organisation. The institution is experiencing a lot of overhead costs to run the everyday operations of the Computer Centre. The Computer Centre is an university unit which provides the institution with Information Technology (IT) services such as Internet resources, systems for processing examinations, salaries, student registration and printing. Other functions include processing teaching, tutorial and research materials. However to accommodate the storage of the university’s huge data, the Centre requires a number of hardware equipment such as mail, SQL, Salaries, examination, human resource and security servers. To maintain the capacity of such data in terabytes under the traditional system is very costly. As the University spent a lot of money to run and maintain just a single server and what more on several of them as listed above. As a result, most of the university’s budget goes to run the services of the Computer Centre. It is against this background that the University envisaged on CRECES, a system whose hardware and software resources are either scaled up or down depending to the size of the workload. The CRECES uses the technologies of elasticity, scalability, virtualisation and Type 2 Hypervisor to produce and increase the number of virtual guest servers during the peak workload period. These resources process the large volume of workload at a very minimal cost of production. Similarly, when the workload reduces, the virtual guest servers are scaled down to a single host machine. In that way the cost of production is reduced drastically hence managing the hardware and software resources at a low cost of production. The CRECES yields a number of benefits that include affordable university operation budget, efficiency in processing of examination, registration, and research data. Others benefits include the capacity building, but to mention a few. Others include upgrading of technician, technology acquisition and capacity building.