A Ten-Step Plan to Effective Capacity Management

Designing and building an effective capacity management function has never been more challenging: the competition for talented staff is high, as is the current workload of projects and 'business-as-usual'. At the same time salary and training budgets are often constrained. How does the contemporary capacity manager cope with these challenges? Follow our ten-step plan to designing and building an effective capacity management function; wherever your starting point and whatever your challenges you will find useful techniques in this article.

The Ten-Step Plan: Overview

The Ten Steps:

1. Organisational Structure
2. Model, Model and Model
3. Align Resources with Service Importance
4. Build a Single Capacity Repository
5. Use the Web for Capacity Reporting
6. Publish Reports on Key Systems
7. Automate Processes Wherever Possible
8. Define Scope of the Capacity Management Function
9. Establish Authority of the Capacity Management Function
10. Audit the Staffs Skills and Define Training Plan

1. Organisational Structure

There should be one centralised capacity management team that covers the proactive capacity planning and reactive capacity management of all ICT platforms within the enterprise. This team should be divided into two functions, with staff ring-fenced and focused only on their function: firstly, project-facing staff (the 'capacity consultants') who deal with all incoming capacity requirements for new and amended services; secondly, the platform experts (the 'capacity service-owners'). Each platform expert should be assigned responsibility for one, or several, end-to-end services and should be cross-trained on the other platforms across which their services traverse.

Additional areas of interest include the need to include performance management (e.g. modelling and tuning) within the same team. If this is not possible organisationally then a strong inter-team relationship should be established.

2. Model, Model and Model

One of the difficulties with offering capacity management to all ICT services within any medium to large organisation is getting the time to undertake a thorough analysis of a service or platform in order to build an accurate capacity plan. The building of a capacity and/or performance model is extremely useful in this situation, as it allows quick analysis of the impact of changes. The time spent modelling of any high or medium priority service invariably saves considerably more time in repetitive analysis when various scenarios are considered.

Ideally models will be built by the experts within a team and then handed over to the junior members to run (as they don't require the experienced staff to do this); they can use the maintenance of the models as a learning experience.

3. Align Resources with Service Importance

ICT has developed to become more mature and aligned with 'the business' in recent years, partly driven by formalisation initiatives such as the IT Infrastructure Library (ITIL), Control Objectives for Information and related Technology (COBIT) and the Capability Maturity Model (CMM). With this as a driver, and the move away from centralised computing platforms (although this is currently being partly reversed with the move towards virtualisation), capacity management functions have had to face a difficult change, moving from resource-based capacity management to service-based capacity management. However, no capacity planning function in any sizable organisation can plan all the services that the business uses; therefore some level of prioritisation-based rationing is needed. The level of resources that a capacity management function should allocate to each service should, generally, be proportional to the importance of the service. This can be determined and achieved in a number of ways - for more information on this particular topic see Prioritising Capacity Management Resources.

4. Build a Single Capacity Repository

A factor which always prevents effective capacity and performance management is the inability to analyse data relating to a single cross-platform service. This is usually not because the data isn't collected, but because it resides in disparate locations and differing formats. The obvious goal is to bring this data into one single database, the Capacity Database (CDB) referred to by ITIL. However this is not always possible nor is it always a worthwhile objective logistically. The objective should not be to get all the data in one place, but to be able to easily retrieve a consistent sub-set of data relating to a particular service or transaction. This objective does not require a single database as such, it only requires a single meta-database containing views which are linked to the underlying data. This immediately negates some common problems that building a single database introduces:

• Building a large central database and the storage of duplicate data - this central database will be enormous
• Transferring the data to a single location - even if undertaken out of core business hours this may impact batch transfers, etc.
• Data integrity issues - if the central database contains incorrect data it is unlikely to be updated

By using a meta-CDB it is possible to run live queries across several databases without transferring the entire data set from each platform, which would contain a significant amount of unnecessary data; this also ensures the data is up to date and data integrity is retained.

5. Use the Web for Capacity Reporting

One area that consistently causes capacity staff resourcing issues is that capacity and performance teams are expected to supply ad-hoc analysis on systems and services, often at short notice, to resolve incidents and problems. This can weigh heavy on the capacity and performance function as it tends to supersede any other work. One method of off-loading this frequently onerous duty is to provide many pre-configured (or 'canned') reports on the intranet. This reduces the time spent on unnecessary reporting. Ideally all reports should be automated and on-line to remove this burden from the capacity and performance staff (who are paid to be analysts, not report writers).

6. Publish Reports on Key Systems

Often capacity and performance functions are only noticed when there are incidents or problems that are blamed on their lack of foresight. This is generally unfair and demoralising for the staff and management of the function. One method of raising the functions profile positively is to provide a regular report on a key cross-platform service. This should contain all the resource data that can be retrieved from as many platforms as possible (e.g. WAN, LAN, server, workstation, etc.) in one consolidated but readable report. Most importantly, in addition to demonstrating the collection of resource metrics the report should contain any service metrics available, such as business transactions, by type, and their response times. This type of report, circulated to the appropriate senior management, will raise the profile of the capacity and performance function positively as it demonstrates a comprehensive, proactive and professional approach to relating to the business.

7. Automate Processes Wherever Possible

Many capacity and performance processes and activities comprise of repetitive work that can easily be automated, such as generation of graphs from data, running capacity models and pulling disparate data together into a single report. It is wasteful for expensive resources, such as capacity and performance staff, to undertake any work that can be automated easily. To automate these processes it is worthwhile considering recruitment of interns, or even temporary staff; if your organisation has an annual graduate intake it would a useful employing a graduate to undertake this automation.

8. Define Scope of the Capacity Management Function

The capacity management function often has a personality crisis over what it is meant to manage. This should be well defined and understood across the entire ICT organisation. Examples of scope issues include:

• Are all technologies included, e.g. TCP/IP routers as well as Mainframes?
• Are all resources included, e.g. data centre power and space?
• Are virtual resources included, e.g. VMWare guest partitions?
• Is application performance included?

9. Establish Authority of the Capacity Management Function

An area where the capacity management function can add considerable value is as part of assessing new software and systems prior to implementation. While application sizing (as it is called in ITIL) is invaluable as a proactive service the capacity management function is not often given any power of sign-off on whether applications or service is 'fit-for-purpose'. Review and approval of new services and systems should be considered as a primary role of the capacity management function, and so it should be given the power of sign-off. Of course any recommendation not to proceed may still be over-ridden by the business when they want, but by providing accurate assessments of the impact of roll-out the capacity management function will gain respect and then authority.

10. Audit the Staffs Skills and Define Training Plan

Capacity and performance staff require a very specific technical skill set and an appreciation of how business relates to the ICT services it uses. Examples include: statistical analysis, queuing theory, trending, etc. These skills are the mandatory minimum and precede any knowledge of individual platforms or technology. Staff should be audited to ensure they have the relevant skills for their roles, and level of seniority within the function, and training plans put in place where gaps exist. This should be coordinated with any defined career path or succession plan within the function.

Summary

A capacity manager does not need to implement all of these suggestions to make a significant difference to their function's effectiveness, any one will make improvements, but the more that can be achieved the greater the improvement.

This article is summarised from the Capacitas presentation 'A Ten-Step Plan to Effective Capacity Management', which was presented at the 2006 UKCMG Annual Conference.

Access to Capacitas articles is unrestricted although research is restricted to registered users of this website; registration is free and available to all. Click here to sign up now. Subscribers will be informed via email when new research is published.