Setting: A bank needs to assign the cost of its IT capacity to a portion of its transactions, but activity levels during its peak-time hours (9:00 AM - 5:00 PM) and off-peak hours are very different. To perform every off-peak transaction with an acceptable level of service, the bank requires a computing capacity of just 100 million instructions per second (MIPS). The same level of service for peak-time transactions requires 300 MIPS of computing capacity, or the current capacity of the bank's servers. The daily cost of capacity for the bank is $12 per MIPS, or $3,600 overall ($12 per MIPS x 300 MIPS). There are no economies or diseconomies of scale for server capacity cost. The server capacity has to be purchased in advance, and cannot be scaled up or down on short notice. First, the bank wants to know the cost of the three transactions - balance checks, fund transfers, and printed statements - during peak and non-peak hours. Balance checks require 1 MIPS, funds transfers require 5 MIPS, and printed statements require 15 MIPS. (Note: If total capacity is 100 MIPS and a transaction consumes 1 MIPS, that transaction can only be performed 100 times in a given hour. In the course of a day, the computer system could perform 2,400 [24 x 100] of those transactions.) Second, assume that the bank wants to calculate divisional P&L on a daily basis. In order to compute divisional P&L, the bank needs to know how much IT expense it should charge to its Individual and Corporate Banking divisions. On this particular day, the server log indicates that the Individual division consumed 1,920 MIPS in peak hours and 1,280 in non-peak hours, while the Corporate division consumed 480 and 320 MIPS, respectively. Questions: A1. Calculate the cost per MIPS during (a) peak time and (b) off-peak hours A2. Using the costs above, calculate the IT cost of a balance check, a fund transfer, and a printed statement during the peak and non-peak periods. A3. How might the allocation strategy you developed potentially lead to changes in the actual cost structure? A4. Calculate the daily IT cost of the Individual and Corporate divisions. A5. How would your calculations for A4 change in a day where the level of capacity remained the same, but the server log indicates that the Individual division has consumed 640 MIPS in peak hours and 1,024 in non-peak hours, while the Corporate division has consumed 160 MIPS in peak hours and 256 MIPS in non-peak hours? Exhibit 6 Calculation of IT Allocations Step 1. Calculation of capacity per time bracket The % of various capacity component can be calculated based on STech (MUIR) data In productive capacity, all users will share costs based on actual consumption and premium will be assigned to peak-hour users Total Practical Capacity 7,000 Planned (Forecasted) 6,000 4B Management Policy Strategic Capacity: and Start-Up 5,000 4A Peak Reserve (SLA) 4,000- 3 Planned, yet Unsold 3,000- 2,000 1,000 0 T. Afternoon Peak 9:45 a.m. Peak (a 15-min peak) T Off-Peak T After-Market Productive Capacity 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 Time . Cost Driver (business transactions) . Get Balance & Get Positions are the top 2 activities, regardless what time of the day (24 hours) Peak load (9:30-9:45 a.m.) is normally 30% to 50% higher than Off-Peak load (10 a.m. -3 p.m. EST) Productive component (area 1) represents near 30% of [Total Capacity minus Strategic Capacity & Unsold and Standby (area 3, 4A, and 4B)] . Definitions: Rn Rate for nth period usage Tn Normalized CPU usage for period n Dn= Difference (delta) between Tn and T(n-1) for period n Cn = Un-normalized CPU usage for period n T, T, and T, are established in the budgeting process T, T, T and T, are computed monthly as the average of the usage for each time bracket in 15-minute intervals T= 3,200 C-3,200 x1 = 3,200 D 200- T=3,000 D-500 T=2,500 C=2,500 x 4.5=11.250 D=2,000 C, 17.5 x 500-8,750 T-500 D-500 43 Zone Duration 11 C=3,000 x1 = 3,000 2 4.5 Morning Peak 1 17.5