The history of computing is well-documented, as is the is age old adage that those who forget the mistakes of history are destined to repeat them.
Computing as a tool to enable a business to manage vast numbers of transactions, (some complex, some simple, always incredibly repetitive and absolutely consistently) is one of the most fundamental functions of computing in business, science, commerce, creativity - in fact, everywhere. Being able to think of new ways of delivering information, services or products has become key to business fitness. Making a change to a complex system quickly is of course a double-edged sword, the full impacts of change need to be assessed, tested and considered before they can be implemented.
The philosophy of delivering rock solid reliability demands incredibly robust processes that drive down to the raw physics and chemistry of an environment to build stable foundations designed to exceed a very clearly defined set of needs. Look at the process of building a large skyscraper, large airplane or really any huge machine. Every element has to be considered to ensure the whole system will work flawlessly for the whole of its designed life and (hopefully beyond).
This same level of engineering control is followed by the engineering team that implements and runs all the changeable systems in these large machines. A skyscraper is continually being reconfigured as different businesses move in, move out, staffing levels change, usage changes etc, but all the subsystems for power, lighting, air, elevators, telecoms, security, fire control, interior design etc. are designed to deliver 100% reliably in this environment of continuous change.
Imagine a building, which didn’t follow these rules! Imagine having for example an open plot of land where different buildings could be quickly built to meet short term goals, where each building on the plot could make their own decisions on security, fire prevention, telecoms , interconnection etc. It would be a fun place to work (at least initially), but keeping control would be a lot more complex, and the likelihood of a problem would be much greater. Places just like this do exist, and they work well for smaller businesses, where higher levels of risk can be taken because the impact of failure is less critical.
I believe this analogy maps well to the world of computing. Windows, and Unix environments allow for quick implementations but by their very nature they assume that all the processes will be used by all applications and all users, and it's up to the implementation and maintenance teams to lock everything down, and consider each and every interconnect to each and every system. Of course there are great tools available to help with the process. But you start from a position of vulnerability and have to be very careful to consider every possible event and condition. In this ecosystem we have seen the development of a whole industry of people who love to discover and exploit vulnerabilities using ever more creative and complex methods.
The mainframe conversely is architected (like the skyscraper) with integrated subsystems in place that assume that by default every application and every user has no rights to any subsystem without it being explicitly granted. A model where every change is modeled tested and the impact considered before being allowed to impact the live system. This leads to very predictable performance, reliability, resilience, and cost control. When everything is considered holistically the costs of the whole is known. In the world of mainframe it is not possible to hide costs within other budgets, while in the world of UNIX and Windows this is absolutely the normal. Just consider the ability of every department to purchase laptops, desktops, software, people, services etc. Every one of these elements is known in the mainframe world, and is only partly known in the distributed world. Today very few distributed environments are anywhere close to 100% compliance to their internal policies of device usage, software usage, configuration, or external access to internal systems. There are just too many ways users can circumvent the rules and any number of these can be exploited either intentionally or unintentionally. Some incredibly clever management tools are available to help lock down the environment, but this takes additional resources to implement, while the mainframe continues to have absolute control over all these aspects.
If the business wants to implement a new business process on the mainframe, it may seem to be very cumbersome, and may seem so much easier in the distributed world. But if the full costs including those of security, reliability are performance of the whole system are considered then the mainframe wins. This glasshouse versus distributed architecture model is not new of course, but a new member has been added to the team, Mr. Governance. Now all business processes need to follow very controlled guidelines to ensure that the letter and spirit of these new laws designed for the age of IT are followed. The costs of not following these rules are high today and are only likely to get higher. Huge fines, jail time and eventually businesses losing the rights to trade are the consequences of not meeting the law. Today we accept that a rat infested restaurant will be closed down, well get ready for the non-compliant business being closed down.
This maturity in the law brings with it the very real potential of an economic infrastructure that treats electronic transactions with the same legal value as paper transactions. This is incredibly good news for business (and is long overdue), but can only work when business is forced to adopt strong governance throughout the whole value chain. It’s happening right now, and this is great news for businesses that are already running mainframes, they already have the basis of exactly the right environment for the next stage of ITization of business. Having a well-managed mainframe is one of the strongest competitive advantages a company can have in the IT governance era.
Key to all this is “well managed” and the leader in managing IT is CA, who today simplify and unify the process of managing , governing and securing IT on the mainframe and distributed IT world.
David Liff