Sep
27
The $’s and Sense of Cloud Scalability
In today’s world there are two basic models for cloud computing: One is utilizing a particular application for a particular purpose, SaaS; two is the process of developing applications with the specific purpose to run an app engine, IaaS. Now don’t get me wrong, both of these service models have a wealth of positives about them, but they are inherently limited by definition. What you are essentially doing here is creating building blocks and placing those blocks in an environment specifically designed for them to succeed. The blocks do what you intended for them and the environment thrives because of it. What limits the services is the lack of flexibility of the blocks. They are extremely rigid and stubborn applications which, in the end, are forever constrained to their singular task in their singular environment.
The progression forward when dealing with this inflexibility issue is to get away from a predefined solution and move towards raw computing power, and that is what Google Compute Engine is. In the past 4 years, Google has begun a movement away from traditional server sellers such as IBM, HP and Dell and positioned itself as the world’s fifth largest server maker. The cloud-based, elastic environment Google created has the potential to be best friend to everyone from garage startup to enterprise corporation:
Example #1: Steady State Organization
A technical solution company has been around for 4 years now and is well established. Their clientele has grown at a steady rate and in turn their number of employees has reached 50. They have a discovered that at their peak work hours they will need about 100 servers to support their entire business model. They had the capital, so they simply went out and bought the servers. What management didn’t consider is the servers rarely, if ever, run at 100 percent. According to Gartner research, Server utilization is often at the low end of the performance range, averaging between 7 and 15% of their potential power. In addition, they are only performing account process for about 3 hours each night, and sit dormant the remainder of the time.
The servers filled an immediate need and the company could afford it, so they can live with the down-time of the servers. What they fail to understand is that by moving these servers into Google’s Compute Engine they would have had a much more dynamic conversion of capital expenses into growth profit. The cloud enables a firm to pay only for the servers while they are using them. The servers would stretch out to the power of 100 servers during peak work hours when everyone is in the office, then scale back down as people begin logging off at the end of the day. By doing so capital IT expense is dramatically decreased.
Example #2: Startup Organization
A Government agency is following the growing trend towards e-government and online service for citizens. Their initial intention was to start online conveyancing and transactions of property sales. In the old world it would take a 5 year, 75 million dollar binding contract to accomplish the goal with an IaaS solution. Again there would be the inherent risk that would be taken on by either the agency or a service provider, lets say IBM. With a million transactions per year, the agency will pay IBM just 15 dollars per transaction. Pretty sweet deal right? Wrong. The deal may seem advantageous at first glance but, this same function could be created in the cloud for one-hundredth of the cost.
Furthermore they cannot easily repurpose that infrastructure when they decide to cut ties with IBM and cancel the project. In all likelihood they could end up having to pay IBM even more money just to come and take the servers away for them! This far too much risk for an organization in the public sector to endure.
If they take their same business plan forward into the cloud model the risk immediately vanishes. They can start with 1 transaction that requires about 1 hour of processing, and thats all they will have to pay for. From there the number of transactions/processing can build and the servers will expand and retract as needed. On top of that, they are free to walk away from the infrastructure whenever they want, with no cancellation fees or useless, space-consuming servers lying around. Further, speculative startups that are unsure of their business model can confidently use cloud based IaaS to grow as the opportunity grows. The engines will scale up or down as your business needs dictate.
Example #3: Need to Start then Eventually Shut Down
The need to begin a project that has a designated and limited run time is not uncommon. Whether you’re business is managing a transition from analog to digital TV broadcasts or running a promotional website for the Summer Olympic Games, the need to quickly create and easily shut down a process is an absolute must. When you are managing a project like this with Google Compute Engine, there are a number of glowing qualities that will enable a smooth and successful project. During the development of the necessary infrastructure you can utilize the cloud to test interim and potential platforms. You can move specific testing environments into the cloud, without transferring the entire business over. This allows for something of a trial-and-error process in designing your platform. If an idea is created in the cloud then deemed unsuccessful there will be absolutely zero decommissioning costs and you can simply wipe the slate clean and start again.
Now if this situation is carried out in the old world IaaS, it is a completely different story. If, after a month of developing a new website you find that there is more work to be done than originally predicted and you’re in need of increased computing power, you will be in the position of ordering say 6 more servers to support the growing project. Those servers will probably ship in from HP in about 6-12 weeks. In the cloud a request for more servers can be filled in about 2 minutes. This ability to make smart and quick decisions about infrastructure is simply revolutionary.
So back to the “shut down” part of this example.. The Olympic Games are over, you no longer have need of the servers and it’s time to shut down this successful project. Instead of being stuck with unused servers taking up space in a data center, or trying to sell them to another department, you now have the pleasure of walking away from the infrastructure contract when you are done with it. No clean up crews, no cancellation fees. You paid for it while you used it, you’re done using it, so you’re done paying for it. No longer do you have to watch the profits of a successful six month promotional campaign dwindle away as you pay the IT costs involved with making the whole thing happen in the first place.
Conclusion
Google Compute Engine has the speed, agility, elasticity and web scale that businesses of all shapes and sizes absolutely need in today’s digital world. It puts IT in a much more responsive position with less capital risk.
Written by Tony Evans, Managing Director of Australia and New Zealand, White Stratus
For more information visit www.whitestratus.com