Grounded in the Cloud
cancel
Showing results for 
Search instead for 
Did you mean: 

Plan infrastructure capacity needs based on biz metrics: No Sweat!

Harmeet_Singh

Every business has peaks and lows in their business activity during a year. Cloud computing has given the choice of adjusting opex in IT infrastructure based business cycles. However this requires planning and knowledge of how much of a resource should be optimum to handle this seasonality in the business.

 

Cloud Optimizer 3.0 gives you the information about how much of memory and compute would be required for a peak or low of particular business metric based on historic data. For example number of transactions happening on a group of VMs can be captured and then fed into the "Business Metric Analyzer" for that the group of VM's running that load. A sample CSV can be downloaded from "Business Metric Analyzer" page for reference.

 

Business Metric Analyzer analyses infrastructure data, and business metric data that can be fed additionally into the system, and derives a correlation behavior between the provided business metric for the same time duration against the infrastructure metrics it has already collected. If there is some significant correlation exists between business and infrastructure metrics, a graph is plotted for business metric against infrastructure metrics. Statistical correlation is a term which tells of how much of a variation in one metric can be explained by changes in another metric.

 

Once the pattern of change between the metrics is established, it is possible to use one to predict the other. This is exactly what is done in Cloud Optimizer's BMA (Biz metric analyzer) module.

 

In this example I have a sample payment application which is running on two subgroups of VM namely “Transactions” and “Database”.  I have collected “NumberOfTransactions” business metric for “Transaction” group for a time period of about 1 day with collection interval of 5 mins. Cloud Optimizer has already captured the infrastructure metrices for the same period. After uploading the business metric CSV file, I hit the “NumberOfTransactions” metric button to see if there is any significant correlation that exists.

CO1.jpg

 

As you can see in this figure above, Business Metric Analyzer has found the business metric and the infrastructure metrics to have a strong correlation. Based on the data from previous business activity peak, the business metric value(NumberOfTransactions in this case) is fed into step 3, for which I want to know how much of resources would be optimum. In this case I am expecting number of transactions to be 3000 during the peak of by business activity for which "Business Metric Analyzer" is suggesting to have 8 CPUS and total compute of 16GHZ with same amount of memory in the group.

CO2.png

With this information I can plan for subscribing extra compute resources during the peak of my business activity much more optimally. The current analysis is based on linear regression which gives reasonably accurate prediction based of near historic(last 1-2 weeks) business metric data.

Cloud Optimizer is available for free and full-function evaluation for 60 days, and permanently free for managing a small environment of less than 25 instances (hosts + VMs). Do try it out and pass on your feedback about how it helps you manage your datacenter (town) better.

CO3.png

Here's a great overview video of Cloud Optimizer's capabilities and functions. 

You can also see demonstrations and find out more features of HPE Cloud Optimizer and related solutions in our sessions and at booths during HPE Discover Las Vegas. Click on the image below or here to register.

CO4.jpg

HPE Software Presales Solution Expert for ITOM
  • HPE Cloud
About the Author

Harmeet_Singh

Currently working as ITOM Presales Solution Expert for APJ. Previously worked with Cloud Optimizer and Operations agent RnD teams as an architect.

Events
28-30 November
Madrid, Spain
Discover 2017 Madrid
Join us for Hewlett Packard Enterprise Discover 2017 Madrid, taking place 28-30 November at the Feria de Madrid Convention Center
Read more
See posts for dates
Online
HPE Webinars - 2017
Find out about this year's live broadcasts and on-demand webinars.
Read more
View all