Fitness for Purpose (FFP)
* FFP is a check on whether the final installation will provide the services or solutions wanted. Most of the time, FFP refers to a set of specific applications related to the business of the organization.
* Speed of computing or communicating- this involves the latency and throughput of the user device, the server device and the connecting fabric. Latency is the wait time. Throughput is the data volume flow.
* Security of computing- this includes the integrity of data, authenticity of parties involved, protection against theft or disruption, and availability of data when needed
* Availability of resources- can computing resources grow or reduce in synchronization with the demands of the organization over time?
* Environmental Protection Aspects such as power consumption, space utilisation, noise level, electromagnetic interference, and use of hazardous materials.
* Each organisation has its specific business agenda and personality. One answer does not fit all. In fact, one organization may come up with a multitude of plausible options. This paper does not attempt to be prescriptive beyond the fundamentals and high level concepts. Consultants will be needed to provide interpretation and advice for specific applications.
Total Cost of Ownership (TCO)
* Cost should not be seen as solely the purchase price for products as many people think and do. Cost should be a bag consisting of hardware, software, consulting, installation, operation, maintenance, migration, outage, opportunity loss, and frustrations.
* We should always set a period such as 3 to 7 years for assessing TCO.
* We should always assess proposals on FFP before TCO, and continue on an iteration manner until we get to a sweet spot on both.
* Some costs may vary during the period. Be aware to factor in cost change.
Hardware Virtualisation
* A main purpose is better utilisation of physical assets. That is, use one physical server to perform as two or more servers. Countering the saving includes the extra cost of virtualisation software, higher level of expertise in installation and maintaining, lower speed performance or response to users, and all eggs in one basket.
* Another purpose is recovery from loss of hardware. Virtualisation allows hardware to be exchangeable and therefore the loss of one would not take much time to resume operation. Users will benefit. If their desktop has died, they can use another desktop hardware to continue. Countering this aspect is the availability of various recovery techniques that may produce similar effects but for a lower cost of ownership. Besides, some hardware vendors produce very reliable hardware!
Cloud Computing
* Cloud computing is also known as utility computing. The word utility comes from the analogy to water, gas, phone, and electricity which come to our home and office as a socket or tap. We turn on the tap and we have water. We plug into the socket and we have electricity. We dial the phone and we are connected. Similarly, we press a switch and we have bits and bytes and are connected to the world. This sounds appealing.
* The data of our organization is unique to us and in many situations is the most important asset of the organisation for competitiveness and productivity. Data is not water, gas, phone or electricity. Data is certainly not a commodity.
* Our computer hardware and software are akin to electricity power generation stations in the sense that they produce the product. The difference is in the uniqueness or commodity nature of the product.
* Data has formats and in fact many formats at present. The formats are not always interchangeable and some are proprietary. This means our data is tied to the software that generates it. If the cloud provides the software, the cloud controls our data. We do not want this!
* Is centralised electricity generation really the way to go? It has happened so in the industrial age. It has happened because electricity generation is not within the reach of an average citizen. On the other hand, computers have arrived on the desktop via the PC magic and the price drops everyday. Industrial Age thinking may not apply to Information Age environments.
* Many European countries are seeing the revival of small scale local heat and power generation. Is this phenomenon rewinding the clock? It seems so. There will be no more transmission grid based blackout threat. There will be no copper loss or energy loss for transmission. There will be no monopolisation and price hikes. These advantages are applicable to the data industry.
Rainforests
* We can tell that Hardware Virtualisation and Cloud Computing have merits and demerits. This is good because we can take the merits and avoid the demerits. We can incorporate both into our Rainforest model.
* A simple planning visualisation technique for managers is the matrix of 4 squares. The 1st square is “Equipment and Expertise In House”. The 4th square is “Equipment and Expertise External”- it is the Cloud. The 2nd and 3rd squares have some internal and some external.
* The Rainforest model helps us put things into perspective for evaluating our options. Always remember the following criteria: fitness for purpose and total cost of ownership. We can treat ownership of data as a fitness for purpose.
Summary
* We should always plan on FFP, TCO, and be aware of changes that will affect FFP and TCO during the period
* We should keep ourselves aware of the position to take on the Rainforest – Cloud Matrix and consider migration options
* Virtualization can be deployed for Rainforests in parallel or isolation to Clouds, but we should be aware that Virtualization is not for everyone.