On June22, 2011, I presented at the Third Annual Cloud Computing World Forum in London. Anestimated 5,000 attendees, over 120 speakers, and 200 private and publicvendors participated in this forum that featured all of the key players withinthe Cloud Computing (CC) and Software as a Service (SaaS) industries. I enjoyed the opportunity to represent ourAgency at the forum, and I presented Cloud as a Game-Changing Business Strategywith the mindset that if used effectively, it will be able to reduce the costand schedule of various projects’ life cycles.
As a highlytechnical, governmental agency, NASA’s success leans heavily on its computingcapabilities. Utilizing the mostefficient and effective approaches for storage, processing, and bandwidth isimperative to ensure our continuing.
Mission-basedprojects within our agency often have very long IT life cycles, and they arealso usually one of a kind and complex: high-stake endeavors. Decisionsfor missions begin early in the strategic planning phase. Due to long lead times for procurementof hardware, IT acquisitions also happen early in the life cycle. Next, because of the long life cycle, upgradesneed to occur regularly; otherwise, what was new when the planning began willbe obsolete by the actual launch date. In addition, the necessary compute must be estimated far inadvance of when it will actually be needed, which can hinder theinteroperability of a project’s components. Another important aspect is that duplicate IT environmentsare sometimes created from the variety of projects (e.g., duplicative developmentalenvironments, integration environments, and test environments) resulting also inthe duplication of software, hardware, and licenses. Furthermore, additional certifications are necessary for anyduplicative environment. These areobvious drawbacks of our current system. However, CC, when strategically used, offers solutions to many of thesescenarios. With the availabilityof on-demand access to compute and incredible scalability (elasticity), thereis no need for advanced estimating, duplicative environments, non-sharing ofexpensive licensing, and non-sharing of HW and SW.
The abovesaid, CC is not a one-size-fits-all solution, because some systems (e.g.,embedded systems) do not fall within CC’s use cases. For example, there is some risk involved when dealing with highlysensitive or complex missions. Also,there are other obstacles, which would need to be considered carefully beforeproceeding: such as theintroduction of possible complications to a project or potential security risks. Additionally, latency can becomean issue once data and applications become distributed and certain spectrums ofcompute (e.g., tightly coupled supercomputing) are needed. Lastly, cultural resistance to change(if disruptive enough to a particular organization) can be a good reason tohold off on a rapid migration to CC.
In orderfor an organization to utilize CC as a game-changing business strategy, itshould be implemented strategically. First, determine where CC can bring the most value to theorganization; next, gauge the levels of potential willingness for adoption. Finally, choose an appropriate approachfor execution: an enterprise-wideor a project-by-project approach. Theenterprise approach was discussed above, so I will address the project-centricapproach next.
For certain organizations, as discussedabove, it may not be feasible to implement CC on an enterprise-wide basis. An alternative then is to use CC on aproject-by-project basis. Such wouldbe possible for six different types of projects.
1. Those that need the latest IT on a “just-in-time” basis. E.g., systemor mission development and construction or renovations of buildings.
2. Projects that need stop-gap capability such as transferring information, consolidating servers, orreorganizing a data center.
3. Risks associated with projects, businesses, and change such as indecisive funding, changes in demand, orsignificant changes to IT.
4. Missions facing challenging funding such as cyclic demands or economic downturns.
5. Improving new projectsto include utilizing CC until the scope is understood or as you need compute.
6. Proofs of concepts. E.g., Scoping of network performance,and scoping workstation or server performance.
In summary,CC should be embedded in the project-management process to ensure business success, risk mitigation, significant cost savings,a better understanding of compute requirements, and the ability to obtain thelatest technologies when buying new hardware. If executed properly, organizations will be able to utilize CCas a game-changing business strategy.