Friday 28 February 2014

Using Systems Capacity Data for Business Intelligence - Creating a plan (7 of 9)

Creating a Plan

I’ve put together the steps you should take to create a plan:
A plan starts by first evaluating where you are currently
Step 1

      •       Who is currently responsible for managing capacity?

       What process and tools are used?

       Is anything being done proactively?

       Are system resources accounted for with costs?

       Is management informed and interested?

       How visible is the activity of capacity management?

       Has the subject been evaluated to determine opportunities?

       What are potential returns on investment (ROI)?

Begin defining processes:
Step 2

       Establish stakeholders

       Identify the customers

       Determine what the customers need and want

       Perform an ROI study

       Gather requirements for what is needed to deliver

       Perform structured analysis on how to deliver

The final step is to implement and deliver, the implementation is driven by requirements:

Step 3

       Determine what software and services are needed

       Formally define deliverables

       Implement software solution

       Architect deliverables with selected software and services

       Automate as much of the process as possible

       Validate operations with review of working environment

       Confirm planned value is being realized by customers

You can use standard project management techniques to help you.

On Wednesday I’ll be looking at implementation and deliverables…..

Dale Feiste

Consultant

Wednesday 26 February 2014

Using Systems Capacity Data for Business Intelligence (5 of 9)


I promised you more sample reports of KPI’s and the next example is storage for IBM SAN controller running on SQL servers.


This information shows metrics bought in as a stacked chart so that you can see the total storage used for this group and you can then use this information for trending, to see when your storage is going to be a problem.

Next I’ve used the same storage example as a trend. I have aggregated all of the sequel servers up and you can see how this will look when you trend it.


This type of information allows us to be proactive as an organization, identifying ahead of time when and where the problems will occur so that we can take action and avoid the issues.

Whilst firefighting may grab people’s attention it is in the best interests of your organization to stop major incidents occurring, and far less costly. It is a much better option to be well planned but remember to ensure that your successes and value remain visible within the organization by reporting on them and making people aware of the consequences that have been averted. Show your value.

Let’s examine where to look for KPIs:

       Business and application transactions

       Transaction response times

       Infrastructure costs

       Budget and planning forecasts

       System and application metrics

       SAN storage metrics

       Network metrics

       Virtualization statistics


       Calculated metrics

These are a few examples but you need to determine what is going to be of most value and importance to you within your organization and look for KPI’s based around those.

Dale Feiste
Consultant

Using Systems Capacity Data for Business Intelligence ( 6 of 9) Costs and Capacity Data

Costs and Capacity Data
A lot of people get very nervous when you start to discuss chargeback, as chargeback can be very complicated and often times is not easily implemented.
You can start simply, without accounting, and it is useful for both you and the users/departments to know exactly who is using what from the total resources.
I’ve given an example below which looks at the top 10 applications and how much of our total storage they are using.
In this instance the large chunk of lilac at the bottom traced back to IT and this is something that you must also factor in – chances are with things like data warehousing and testing IT are heavy consumers in their own right.
Simple chargeback schemes can encourage savings but be sure to include the IT department in your calculations.
Multiple cost tiers for resources can encourage further savings as high end storage carries a much higher cost.
In this scenario our client encouraged users of high end storage to migrate to lower cost storage if it was not really necessary and made significant savings.

Visibility into the cost of consuming resources can be simple
In this example the resource is shown as the turquoise bar chart and the actual cost is shown as the dark blue line.
A simple chart but very effective when talking to your users as you can start to talk in monetary terms about the IT resources that they are consuming.
On Monday I’ll be looking at creating a plan.
Dale Feiste
Consultant

Monday 24 February 2014

Using Systems Capacity Data for Business Intelligence ( 4 of 9)


Key Performance Indicators (KPI’s)

The advent of virtualization has added to the types of system metrics over the years, once upon a time we would never have had to worry about ballooning or dynamic memory balancing. Below are some of the ‘new’ metrics available:

CPU                                                          

·        Avg. CPU Usage in MHz

·        CPU Ready Time

·        Hypervisor Logical Processor %

·        Hypervisor Virtual Processor %


Memory

·        Avg. Memory Usage in KB

·        Balloon KB

·        Swap Used KB

·        Consumed

·        Active

·        Dynamic Memory Balancer

·        Dynamic Memory Pressure

·        VM Vid Partition


I/O

·        Queue Latency

·        Kernel Latency

·        Device Latency

·        Virtual Storage Device

·        Virtual Network Adapter


KPIs can be any time series data that provides valuable information and it is for you to determine which are the most valuable to your organization.

The metric chosen for the chart above is not a system related metric it is a business metric based on revenue, this could just as easily be based on something like transaction count. Bringing in business data gives you the advantage as you can then view business metrics in the context of your system data.
The integrator feature in athene allows you to integrate all sorts of time series data so when you choose your metrics you are not just restricted to basing your KPI’s on pure system data – this allows you to introduce metrics that are also important to your stakeholders.
I’ll be showing you some more example reports on Wednesday.
Dale Feiste
Consultant

Friday 21 February 2014

Using Systems Capacity Data for Business Intelligence(3 of 9)


Evolution and Creation of Capacity Data

Nowadays there is a lot of data out there but I have chosen to show a simple chart which is an example of a traditional high level capacity plan.

Below is an example of the diversity of metrics that are out there today and available to the Capacity Manager.

As you can see there are many different data sources and the challenge for any Capacity Manager is which metrics are important to the business.
With this in mind I’ll be taking a look at Key Performance Indicators on Monday.
Dale Feiste
Consultant
 

Wednesday 19 February 2014

Using Systems Capacity Data for Business Intelligence (2 of 9)

Following on from Monday today I said we’d take a look at the architecture and technologies that you are going to use.

So what technologies are used for capacity management?
Data collection techniques – you are going to need some way of collecting data and bring that data in to a database, in a way in which you are able to utilize it.

Centralized storage and management of data – data needs to be easily accessible and also needs to be managed effectively. You will need to keep large quantities of historical data if you wish to model, trend and chart effectively and if this isn’t managed correctly it can become unwieldy very quickly.
Reporting and alarming facilities – notification is important as you want to know when you may need to do something in your environment. Set thresholds and receive notification when they are breached or about to breach.
Predictive modeling capabilities – forecasting (see bottom right of reports above for an example forecast report), you can use these to predict Service Level Agreement (SLA) requirements, determine what service levels will be in the future and when service levels will be exceeded.
Best practices for Capacity Management
         Use a single repository and tool
         Effectively manage historical data
         Automate where possible
         Utilize proactive notification of events
         Report on data from different perspectives
         Start small and grow
         Ensure realization of value
On Friday we’ll discuss the evolution and creation of capacity data.

Dale Feiste
Consultant

Monday 17 February 2014

Using Systems Capacity Data for Business Intelligence Blog Series (1 of 9)

I don’t think anyone would disagree with me when I say that informed decisions are critical for success, so what is Business Intelligence?

Forrester’s definition of Business Intelligence is:
"Business Intelligence is a set of methodologies, processes, architectures, and technologies that transform raw data into meaningful and useful information used to enable more effective strategic, tactical, and operational insights and decision-making."
So let’s start by splitting that down and begin with a look at methodology and process relating to capacity management.

Methodology and Process
What is a process?
A process is a collection of related, structured activities or tasks that produce a specific service or product, for a particular customer or customers.
What are the first steps in creating a process for capacity management?
      Establish stakeholders
      Identify the customers
      Determine what the customers need and want
      Gather requirements for what is needed to deliver
      Perform a structured analysis on how to deliver

This is really the planning stage, where you are getting this ‘upfront’ information gathered and documented so that you have a good path forward when you go to create the solution.
The next step will be to look at the architecture and technologies that you are going to use and I’ll deal with this on Wednesday.
Dale Feiste
Consultant