Green Data Centres, Virtualisation, Cloud Computing and Unified Threat Management - emerging academic research areas for dissertations and theses

In the modern world, Information and Communication Technologies are very closely integrated to form total solutions for businesses. Hence, many academic topics for dissertation and thesis research projects can comprise of problem areas addressing both these technologies when investigated in the context of corporate business solutions, solutions for government organisations and public infrastructure services. Some of the solutions in IT are widely debated because they are being claimed to be the future of computing infrastructures for IT enabled businesses. I have come across numerous white papers that attempt to establish the feasibility of these technology solutions. These white papers, mostly sponsored by original equipment manufacturers, solution providers and service providers, have been very effective in outlining the benefits of these new solutions and their high level design details such that corporate business owners have started taking interest in them. Virtualization, Cloud Computing, Green Data Centres and Unified Threat Management Solutions are four such areas for which the global vendors and service providers are actively pursuing their customers. Many customers are already running pilots in their IT infrastructure systems but the mechanism of learning from the pilots is not clearly defined and implemented. Such large scale changes in the world of IT systems and networking cannot be implemented based on ad-hoc learning from pilots given that unstructured learning approaches can lead to incorrect and biased conclusions thus causing major setbacks to the businesses. The focus should not be only on saving capital expenditure and operating costs but also on IT services, IT governance, Information Security, Compliance, Reliability, Business Continuity, etc. In fact, the ground level implementation plans and their challenges are yet to be analysed, tested and ratified. The academic community can find numerous opportunities in establishing the validity of these new solutions. The students should focus on studying the realisation of business benefits claimed by the OEMs and Solution Providers such that the other side of picture evolves clearly. I hereby present an outline of these solutions for the benefit of students undertaking higher studies in IT systems.

Green Data Centres: This is also referred to as sustainable data centres by many analysts. The detailed specifications for designing and deploying green data centres have been released by many companies and independent technology analysts. The primary target of green data centres is to achieve "conservation as much as possible" - energy conservation, space conservation, cost conservation, resource conservation, etc. The designers try to implement systems that are as lean as possible. But Gartner reports have warned about threats of crossing conservation thresholds that can result in reduced performance, reduced productivity, reduced disaster recovery capability and above all, reduced capacity and flexibility to take on the business growth challenges. Unfortunately, the consulting world is closely affiliated to the OEMs and Suppliers and hence all designs and solutions are normally biased to achieve sales targets. Hence, I suggest that students should come forward and undertake dissertations and thesis research projects to study the designs, implementation plans and maintenance/running/upgrading challenges of green data centres. A number of topics can evolve especially if the studies are focussed at the local geographies where the students are residing. The reader will appreciate the fact that medium to large scale data centres require enormous power capacities that the buildings meant for office spaces cannot build to host them even if they have the requisite space for hosting the racks for network devices and servers. In my consulting assignments, I have struggled significantly to fit the equipment power ratings into the power budget provided by the building administrators. A medium to large scale data centre may require anywhere between 500 KVA to 1200 KVA (or may be more) of power capacity which is not provided by even large scale builders offering office spaces of the order of 100000 square feet per floor or more; and yes, please keep in mind that I am only talking about the data centre and not the desktops, laptops, lights, airconditioning, heating, etc. of the employee areas. With the rapid growth of businesses in the modern world of globalization, data centres can no longer be squeezed and made a bottleneck for business growth. This is, probably, the last thing upon which a business may like to compromise. Hence, Green Data Centres appear to be the solution for the future. Also, it will be well integrated with the philosophy of the Green Building revolution across the world. Every original equipment manufacturer is working towards reducing the power consumption ratings of its products. They have already done well in reducing the form factor and hence the heat dissipation of servers and network equipment, but energy conservation is still not addressed adequately. Overall, the total solutions should be a combination of energy efficient solutions and products. In this context, virtualization is gaining significant popularity across the world. The students may like to conduct separate researches on implementation of green data centres and virtualization or else combine both of them to conduct integrated researches. 

(b) Virtualization Solutions: Virtualization has become the buzz word when future solutions for IT enabling of businesses are discussed. This refers to the technology in which multiple virtual machines can run on a single hardware as if they are independent computers used by independent users. Some strategists argue that virtualization is one of the key deliverables in deployment of green data centres. The products from Microsoft, VMware, and Red Hat can enable end to end implementation of virtualization solutions. Many companies have already started implementing virtual servers in their data centres hosted on blade server hardware. With all the buzz around, very few have employed structured research procedures to determine whether the self hosted virtualization solutions are able to deliver to the business as per the claims made. I suggest that the students should come forward and employ empirical techniques like Phenomenography to investigate the actual business benefits achieved by corporations by implementing in-house virtualization. A large number of topics can evolve in this problem area. The focus should be on cost reduction as well as improvement in productivity and performance of the business. Gartner reports have warned about many negative effects of virtualization if the corporate strategies, performance objectives and corporate governance/information security objectives are not included in the architectures designed by the solution providers. In the consulting assignments of ETCO India, I have observed that the business stake holders are very reluctant to accept virtualization to host their business critical solutions due to absence of proven track records and absence of empirical generalizations in the academic world. This is a very vast area for academic research. The students can create multiple virtual hosting scenarios on OPNET Modeler or similar tools and simulate the models to generate results. Additionally, the students may like to conduct case studies on Corporations, SMEs and Entrepreneurs that have already hosted their IT systems on virtual servers. The attributes to be investigated are: performance of servers after hosting multiple operating systems and applications, network traffic, user experience, possibility of virtualising IP multimedia solutions, fail over, data organization and management, information security, software delivery and licensing, auditing, IT governance, IT services, competitive advantages gained by businesses, etc.

Cloud Computing: This is evolving as a service facilitating IT resources on demand by virtue of applications and business services hosted on Virtual IT Infrastructures. Many OEMs have already launched cloud computing services to corporations across the world - like IBM Blue Cloud, Google Apps Cloud, Amazon Elastic Compute Cloud (EC2), Microsoft Cloud, etc. These service providers claim that the customers can get any IT resource on demand - storage capacity, memory, network bandwidth, application license, etc. The market is developed to such an extent that millions of customers are already availing these services. The students have significant opportunities to study the benefits of cloud computing to businesses across the world. A large number of case studies is possible because the concept has gained popularity across the globe. It needs to be investigated if the current virtualisation service providers on IT infrastructure clouds are fully ready to undertake the responsibility of running mission critical businesses (like banks, financial services, trading and investments, etc.) the way they have been running reliably in traditional data centres. It will be quite interesting for the students to conduct interviews with professionals that have already hosted their services on virtual servers. The attributes to be investigated are: Reliability, Uptime, Speed and Performance,  Elasticity (resources on demand), Billing, Information Systems Strategy, IT Strategy, Information Security, IT Governance, IT Services (to end users), etc. The cloud computing service providers have emerged into three categories - Software as a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS) (these three categories were originally conceptualised by NIST, US Department of Commerce). The three categories have evolved due to different business models, different ways of customer engagement, and investments. The SaaS providers normally engage with SMEs that do not want to invest on data centres and large scale expensive software applications. They are dependent upon PaaS and IaaS providers. The PaaS providers normally engage with SaaS providers or application service providers (ASPs) for developing, integrating, hosting and maintaining software applications, and are dependent upon the IaaS providers. The IaaS providers normally engage with PaaS providers, and with large enterprises that make significant use of cloud hosting along with their own self hosted virtual infrastructures. The students may normally get access to SMEs and hence can focus on their perspectives of defining what they need and how much they need from the SaaS providers, and what they get from the latter. For example, the SMEs may first design the business requirement specifications and translate them into technology needs (concurrent sessions, MTUs, minimum and maximum file sizes, inter-arrival times and pattern, database connects and queries, storage space for files and databases, retrieval times and frequencies, report generation, DSS transactions, numbers and frequencies of e-mails, concurrent users, etc.). The technology needs can be modelled in OPNET to generate the application demands and the resulting capacity requirements on databases, servers, and networking. The modern concept in cloud computing is pertaining to Cloudlets that can be viewed as service units packaged with multiple components offered by SaaS, PaaS and IaaS service providers. Cloudlets can be modelled and simulated on CloudSim, and also on OPNET and OMNET++ to some extent. The students may like to focus on determining various performance, behavioural, capacity, security, availability, resilience, etc. factors from the perspective of SMEs, and model the ways by which the three types of cloud computing service providers can deliver them. Let us understand this by an example. Suppose that a SME needs 1024 to 4096 bytes of database files to be queried by a browser based application client that fires 10 queries per minute per user. This is needed in parallel with 30 mails per hour per user, and 10 files of sizes 100KB to 400 KB transferred per user per hour. These parameters can be easily configured on OPNET. The resulting statistics of application demands and resource utilisation (database, servers, networking) can give an idea of what is needed from the SaaS provider. The SaaS provider will estimate the capacity of cloudlets based on the statistics and communicate to the PaaS and IaaS providers. The resulting service configurations can be offered to the client at a recurring price. This is a better approach than signing the capacity on demand SLA, in which the budget proposal to the management is impossible before signing the contract. In fact methodology of signing cloud SLAs is one such area where numerous topics can be formulated.

Unified Threat Management (UTM) Solutions: Unified Threat Management services framework is a new innovation in the world of Internet Service Providers using network and host based security products operating on cloud computing platforms. This framework is expected to create new waves of user expectations, service offerings, revenue models and client engagements that have not been tapped till date due to lack of empirical models. The SMEs and Corporations looking forward to transitioning their IT systems to Cloud Computing platforms can hire UTM solutions from an ISP connecting them to the Cloud Computing vendor. One can imagine AOL, AT&T or British Telecom connecting a large client with globally dispersed users to Google Apps through
UTM protected networked links from client desktops/laptops to Google servers whereby all the security controls are taken care of by UTM devices implemented by these ISPs. This is an emerging area that requires enormous research efforts, especially from students. Consolidating security solutions with one service provider has many implications in terms of reliability, dependability, rate of attacks and breaches, third party (service provider) compliance to the information security policies of the customers, ownership of damages to the businesses if things go wrong, strategies to switch service providers, etc. I suggest that students should undertake studies on comparison of UTM solutions with traditional in-house security implementation of corporations from business as well as technological perspectives.

I suggest that students undergoing advanced courses in Information Technology and Communications should develop new topics in these areas and conduct researches for their forthcoming dissertation and thesis research projects. If all the current challenges are brought to the table, I can visualise more than 100 topics on which the students and academic researchers can undertake research assignments. Some of these topics have already been undertaken by students employing our support and mentoring services but more contribution is required from the academic world. Tools like OPNET IT GURU and OMNET++ can be employed to simulate various real life networking solutions to verify the behaviour and performance of these modern technologies in a laboratory environment. I personally like OPNET IT GURU because of its capability of simulating real world wireless products (like Cisco High end switches and routers). OPNET IT GURU academic edition is offered free of cost to students by Opnet Technologies Inc. under their university program. The academic version possesses all the features of OPNET except that it can simulate the maximum of 50 million events which is, however, more than sufficient to simulate any network model created for academic research.