Introduction A TECHNOLOGY COST MODEL FOR SERVER INFRASTRUCTU

Introduction A TECHNOLOGY COST MODEL FOR SERVER INFRASTRUCTURE MANAGEMENT

A TECHNOLOGY COST MODEL

FOR SERVER INFRASTRUCTURE MANAGEMENT

Prepared by Russell A. Rogers, Bank of America

32nd Annual International Conference of the Computer Measurement Group, Inc.

December 3-8, 2006

Reno Nevada USA

Today, a method used for budgeting server growth is linear gross estimation. The method is simple, a manager applies the percentage increase in business growth to the total number of servers they currently have in production. However this method can create significant distortions.

This paper reviews an approach to defining a technology cost model using principles from Activity Based Costing. The model is designed to give managers the data they need to make better decision regarding server budgets.

Introduction

Today, the method many managers use for budgeting server growth is linear gross estimation. The method is simple, a manager applies a percentage increase in business growth to the total number of servers they currently have in production. For example if it is determined that growth in product sales for company XYZ, will increase 10% over the next 12-months, then the increase in server capacity should be 10%. The problem with this method is that it assumes all costs scale in the same way, and it assumes current server utilization is near, or at full capacity. In large enterprises this method can end up inflating server capacity. Managers accept these risks because they lack the information needed to make better choices.

At Bank of America, a pilot project is under way to develop and use a technology cost model that could improve fiduciary responsibility and create better forecasts for server growth. The pilot is utilizing a relatively new accounting technique called, Activity- Based Costing (ABC), to generate a cost per transaction for each product or service. In technology terms, this means defining a model that ties server cost to resource utilization and business transaction volume. The goal of the pilot project is to evaluate how well this new cost model can help business managers plan and grow their server infrastructure, increase average server utilization, and reduce the rate of server growth across the enterprise.

This paper will discuss the cost model developed, and it will discuss how this model is being used today to inform business managers about the cost and the value of their server infrastructure investment.

Activity Based Costing

Developed by Robin Cooper and Robert Kaplan at the Harvard Business School in the 1980s [1,2,3], ABC is an accounting technique that assigns costs to products, or services based on the utilized capacity of resources used to deliver a product or service to a customer. For example, a resource can be a stock trader who buys and sells company stock. ABC can determine the cost related to the process of buying or selling a company stock.

ABC begins by identifying the product or service being sold to a customer. In this case, it is the buying and selling of company stock. Then ABC identifies the key business activity that drives the product or service. Again, in this simple example there are 2-activities, a stock buy and a stock sell.

Introduction A TECHNOLOGY COST MODEL FOR SERVER INFRASTRUCTURE MANAGEMENT

Next, ABC examines the processes or steps required to complete the business activity. In this example, the steps required to buy or sell stock. The next step in ABC is to identify the resources and measure the utilized capacity of each resource. The stock trade takes two types of resources – the trader, and one or more computers. For this example, we assume a buy and sell transaction are essentially the same. ABC determines that it takes the trader 2-minutes to take and place an order. If there are 8-working hours in a day, that translates into 480 minutes per trader. This means one trader can place a maximum of 240 trades per day at 100% full capacity. Say the trader works 240 days per year, and costs the company $200,000 per year in total compensation. If the rate charged by the trader is $6.00 per trade, then at full 100% capacity, the total revenue potential for this one trader is $345,600 per year. At 50% capacity, the trader brings in $172,800 per year in revenue, at about $100,000 in cost, which represents a profit of $72,800. At this point ABC has effectively developed a cost model for the buying and selling of stock. This model computes the cost per trade for the company (based on the time analysis of this single trader) and finds that the cost per trade is $3.47. Now with this analysis ABC can create a summary report that gives a business manager a great deal of useful information to evaluate costs relative to revenue for each product sold – in this case the buying and selling of stock. It also identifies specific cost drivers, and it provides insight into what cost drivers could be optimized to improve overall profitability of the business. In this example, suppose through the analysis of the steps taken to execute a trade, it is determined that some steps in the process could be eliminated, which reduces the time it takes to place an order by 1 minute. This would shrink the cost per trade down to $1.73!

Along with the value of identifying cost drivers and measuring the cost and profitability margins associated with delivering a product or service, the cost per trade model can also be useful in planning business strategy and growth.

If we can measure resource utilization based on current business volume, we can determine how much additional growth in new business could be satisfied with current resources. Any gap between the planned increase in new business volume, and the available capacity to handle future growth, estimates how much additional capacity is required. So, ABC predicts business cost and …… 此处隐藏：5763字，全部文档内容请下载后查看。喜欢就下载吧 ……