by Brenda J. Christie
|Accessing Mainframe Costs|
I recently attended a webinar which centered on the metrics which compare mainframe costs versus server costs. The webinar, “The Surprising Economics of Mainframe Technology“ was hosted by Compuware. One of the keynote speakers was Dr. Howard A. Rubin.
In order to compare the costs associated with a mainframe platform to the costs typically found in a server configuration, Dr. Rubin used several metrics which are not widely known, but which perhaps show a more comprehensive representation of cost for each. Some of the metrics Dr. Rubin used for this purpose appear below. Why any of this matters at all really has to do with declining revenue relative to increasing IT Infrastructure cost. That is to say, Dr. Rubin’s conjecture is that IT costs are rising faster than revenue.
Given the proliferation of technologies (think disruptive technology) and technology driven cultural changes, there is a fair amount of truth to this conjecture. Many people perform work on the way to and from the office. There is a cost associated with company-sponsored cellphones to keep employees in contact with the home office. Companies often provide the devices used by the employee (laptops, tablets, cellphones, and software licenses for laptops, tablets, cellphone and Wi-Fi expenses for these devices). Additionally, staff or services providers to handle maintenance of these devices as well as troubleshoot also contribute to the cost of the mobile employee. Remote networks, software to support remote networks, staff to maintain both, also contribute to rising IT costs.
Consumer demand also drives rising IT costs through commodity availability. Consumers perform banking transactions in a mobile environment. Consumers shop online in a mobile environment. Appointments are made, not by phone, but via the net. Advertising dollars to alert and attract mobile consumers must also be spent lest the consumer stray into competitor territory. Marketing, more than ever, takes place in an electronic environment rather than through traditional paper-based channels.
All these activities drive up IT costs and eat into company revenue. This seemingly insatiable demand fueled by technology is referred to as Technology Intensity – more devices, more apps leading to more demand. Rising IT costs relative to revenue is a direct result of the availability of the Internet, mobile technology and the cultural change whereby employees and consumers are not stationary behind a desk, but on the move. While a company’s revenue may be increasing, the technology, infrastructure and hardware costs eat into that revenue to a larger extent today than in prior decades. Indeed, the actual cost, according to Dr. Rubin’s presentation, was 30 times more expensive in 2010 than it was in 1980!
It would be easy to look at only one side of the equation, i.e., costs associated with the rise of a connected economy. However, if mobile transactions were failing companies, the practice of enabling them would be soon abandoned. The other side of the equation must be examined as well, i.e., rising revenues due to the ability to transact business, make online purchases and stay connected to the office.
Comscore estimated Cyber Monday to be the heaviest day of the 2014 year netting more than $2 billion dollars.
As reported in E-marketer, 41% of online purchases were made using a tablet or smartphone.
Much of this volume involved mainframes. Mainframes support this volume in the background through banking, credit card and inventory applications. They are often also at the forefront through providing analytics to facilitate real-time marketing to further increase sales.
One of the metrics Dr. Rubin discussed illustrates the proportion of revenue supported by mainframes. He applied the same metric, Income Supported per Dollar of Infrastructure Expense, to server configurations as well.
When the mainframe Income Supported per Dollar of Infrastructure Expense (ISPDIE) is compared to server ISPDIE over a period of five years, the result is that mainframe supported more ISPDIE while showing a 76% decrease in cost over the same five year period while server heavy configurations actually supported less ISPDIE than at the start of the five year period.
Part of the reason for this has to do with the difference in how the two configurations scale. Scaling on a mainframe can be accomplished through a variety of means, most of which do not entail buying a new mainframe. Storage may be added, software such as DB2 Analyzer or Sub System Optimizer may be employed to effect efficiencies.
Scaling on a server, however, usually is accomplished through the purchase of additional servers which are themselves accompanied by increased licensing costs, increased labor cost, storage, power and cooling costs, to name just a few. Indeed, there is an adverse relationship between mainframe unit cost and workload. The higher the workload the mainframe processes, the lower the unit cost. The same cannot be said for servers. Increased server workload usually results in increased cost through the items mentioned previously (more servers, storage, power and cooling costs, etc.).
In summary, the growth in technology-driven demand has not only absorbed a greater percent of a company’s budget; cost associated with that demand outpaces the rate of revenue growth. Use of mainframes both facilitates that growth and works to mitigate the cost associated with supporting that growth. This is readily visible when viewed through a new metric, Income Supported per Dollar of Infrastructure Expense. It can also be seen when looking at IT Cost of Goods across different sectors wherein the infrastructure cost used to produce a good or service is examined. That service could be a banking transaction, a credit card transaction; it could be the cost of transporting a person from point A to point B. Across most sectors, according to Dr. Rubin’s research, that cost is less when mainframe technology is used.