Sarquol Ltd

The following story illustrates why capacity management is needed to consider the most appropriate action to improve performance. I suspect you could insert any resource you wish in place of “network” and the assertion would still be valid.

Firms waste billions on network over-design

Who needs desktop Gigabit Ethernet anyway?

http://www.theregister.co.uk/2006/05/22/network_overdesign_gartner/

It is true that simply upgrading hardware until the system performs well can look like an effective solution in many cases. It can avoid often complicated analysis and testing work, and in many cases will be effective in the short term. Simply continuing to throw hardware at problems as a strategy, however, doesn’t really work. The issue is that solving a solution in this manner will generally be a short term fix. If the solution continues to be to keep adding more resources then eventually the costs stack up to be unacceptable. In many cases there will also be a point of diminishing returns where even the best hardware on the market will fail to keep up. A more informed decision can be made by characterising a system and estimating the capacity that may be required to meet future requirements.

When implementing a new system it is common to make high level assumptions about how the system will be used, based on the information provided by the business community on how they expect the system to be used. This form of input tends to be fairly high level, such as the number of transactions per day that they expect once the system is fully operational. This is a strong start point for working out the likely level of demand that will be placed on the system, which is a process that involves a set of estimates and inferred values in order to arrive at a likely behaviour profile. These figures are then used as the basis for performance and load testing of the system. Consider the future briefly though, since this is also when a little forethought in the specification and design of the original system can deliver a huge benefit. Read the rest of this entry »

A second item that I have put together for use on assignments is a “Generic Performance Model”. This model is built as a large Excel Spreadsheet (20Mb download). Using it allows examination of how system performance may vary with user demand, time and system calibration parameters. Since it is a large spreadsheet it takes a while to re-calculate when changes are made, but it does allow useful evaluation of the likely behaviour of a system. Some caution is needed in using it as the usual IT principle of “Garbage in, garbage out” applies. I will provide documentation of the various pages of the model over the next few bulletins, or please feel free to join my web site and ask questions there.

The model may be downloaded from: Generic Performance model

Please Note: The spreadsheet comes for use “as is”. If I find, or am informed about, defects I will resolve them – then  provide feedback in the forums of my web site and in the next monthly bulletin. It is, however, a complex spreadsheet which I am providing for free and so can not provide a guarantee beyond this.

The level of performance proof required of projects can fall on both sides of an optimal “cost – benefit” line for a system. If a project spends too little effort on proving performance then the result will be problems and complaints in production. If a project has overly stringent performance criteria, the result will be high costs and long delivery delays for the project. Read the rest of this entry »

It would seem that “Service Oriented Architecture” (SOA) is taking off as a technology in fashion. As usual, most of the news is generated by people wanting to sell it as the next big thing ready for the prime time. If followed as it is being sold at present, the concept is liable to lead to significant performance problems. SOA is sold on having a large number of “users” who are then often other systems. This is followed all the way back to the real users through an unknown number of tiers. The complexity of managing this sort of architecture increases significantly as the number of tiers increases. If this is then managed without due consideration for managing the system capacity and performance the result will be problems in the future. That is not to say that I don’t see a benefit in the overall technology – there is great potential. Just consider carefully how the performance of operational systems might be proven to be sufficient for the long term.

In a previous employment, the employer had an operational department who were responsible for signing off the performance of a system before it was allowed in to production. Their method for doing this was working with delivery projects to make sure that the project provided sufficient evidence that the system they were delivering would perform in the long term. They would also check that the operational managers had access to appropriate mechanisms to monitor that the systems were performing according to the evidence that had been provided. When requested to do so they would provide expertise to the delivery projects in the appropriate use of capacity management techniques, but they were primarily a gate keeping and monitoring function. Read the rest of this entry »