According to Gartner’s research, two of the most common reasons why an organisation chooses to migrate their application and data hosting infrastructure to a cloud model are to reduce costs and to improve performance.
In theory, the cloud offers clear cost benefits compared to legacy data centres: Your organisation is no longer directly responsible for procuring the hardware that apps will run on, and many of the day to day costs such as power and maintenance are absorbed into the overall subscription costs.
The ability to provision a wide range of compute and storage resource within Amazon Web Services provides a great deal of flexibility when it comes to creating the foundation for your applications to run on, and by tailoring the architecture of your cloud deployment to the specific requirements of your application, performance will be optimised for end users.
This said, why is it that the cost of virtualisation can increase over time while performance can fall below expectations?
Following a move to the cloud, many companies will see their monthly subscription costs increase, seemingly without explanation, while performance can stutter, with apps running at inconsistent levels throughout the day. Without a clear understanding of their cloud environment, it can be frustrating to those organisations, and diagnosing the underlying issues can be nigh on impossible.
A cloud deployment may consist of multiple virtualised systems handling different functions of a typical application. Front end web servers handle the user interface, database and storage systems handle the back end, and there may be individual machines configured to handle specific functions such as identity management or indexing of data. Each of these machines will be deployed to a specification that should be suitable for its use.
Putting in place a monitoring solution alongside your cloud deployment provides an additional layer of information that helps diagnose issues.
By analysing the load on machines over time, it is possible to identify periods of peak demand. During these times, more users may be accessing data at the same time or using functionality that involves higher than normal levels of compute. During these periods, when machines are running at levels above their design spec, bottlenecks will be created that will slow down requests from other users.
Our Cloud Control system monitors usage at all times across all your cloud resource to identify periods where it is under strain. Once analysed, this data can be used to create a more flexible environment where additional resource – or more powerful machines – are specified to run at times when the additional power is needed to remove the bottlenecks. This can be scaled back when not needed to minimise cost increases.
The cost of your AWS subscription is calculated from the number and type of machines that you have in your deployment along with the amount of data that is being managed. While this is straightforward, the billing processes are not. A lack of clarity in your invoicing means that while you have the convenience of a single line item on your balance sheet, Amazon do not provide you with contextual information about what you are actually paying for.
To resolve this issue for our clients, we developed a system called IT Cost Control. This provides transparent information about your whole deployment including data about usage levels. Monitoring data from Cloud Control, combined with the rich billing information from IT Cost Control provides a detailed picture of your cloud consumption, enabling our consultants to audit your usage, identify waste, and streamline your infrastructure so that you only pay for what you need. In some cases, this approach can reduce ongoing subscription costs for our clients by up to 30%.
Both IT Cost Control and Cloud Control are offered as part of our standard managed services packages for both Azure and Amazon Cloud, meaning that all of our clients benefit from a fully monitored cloud solution for their application which gives improved end user performance and lower costs.