Apr 8 ’10
IT Insight: The Mainframe in the Cloud
When the subject of cloud computing is raised, there’s an implicit assumption there are racks of Wintel servers inside the cloud. Why is this? For a start-up, the upfront costs associated with these servers may make them appear attractive, but as the cloud grows and becomes more pervasive, the cost of the hardware isn’t the only issue that matters. Since users of cloud resources have no idea what’s inside the cloud, there should be no particular user preference for one platform over another. But for service providers of public or community clouds, or administrators of private clouds, the price of the server or system should be only one aspect of their assessment; even Total Cost of Ownership (TCO) is only one key factor for prudent decision-making. For all factors, System z is the leading platform for cloud computing.
Availability and reliability are vital to business survival, and no other environment can command the near-100 percent availability of the mainframe. Downtime can result in lost business as well as damage to your corporate brand or organization’s reputation. Resilience also is a factor. Mainframers pioneered virtualization many years ago and have refined this capability to where resources can be allocated and reallocated on the fly. This is invaluable in managing the diversity of demand experienced in a cloud environment. In addition, the mainframe’s failover capabilities are unmatched by other platforms.
One critical area of concern in the current cloud world is security. Mainframes have always been known for their superb security and continually enhance this capability. While it’s possible to fence off users from each other, as outsourcers have been doing for years, the data in a shared cloud can be as secure as your business requires. This might obviate concerns about working in a shared cloud and support the desire of business units to fence off work, if desired, in a private cloud.
Price is a huge issue for both businesses and public organizations. Although the start-up cost of a small server appears to be the best bet, the TCO for mainframes is significantly better in price/performance. IBM has continued to reduce mainframe ownership costs in these areas: software with sub-capacity pricing; hardware with specialty engines (which also may contribute to lower software pricing); people (a large mainframe can manage thousands of applications with only a few staff supporting them); and power (mainframes require less power per processing unit and cost significantly less to cool than distributed systems).
The mainframe was the innovative platform for virtualization, so it’s no surprise it also supports all of virtualization’s enabling technologies, such as Web services and Service-Oriented Architecture (SOA). Though a very complex system under the covers, the interfaces are increasingly simple and designed for people more accustomed to working on other platforms. As with our interactions at an ATM, consumers don’t need to understand what’s inside these systems to benefit from them. This must be true in the cloud world for it to work as predicted.
There may be specific requirements for other servers to be part of your cloud environment, but as you and your management teams plan to migrate some or most of your internal and external operations to cloud computing, mainframes should be considered as the “heart” of your cloud environment. When you think of all the key factors you would want to use to determine your internal or external cloud architecture, the “Seven Virtues” of the mainframe (as highlighted below) at a minimum align closely to this list.
Mainframes defined rigor in security, change management, availability, and increasingly in manageability. So, it stands to reason they also will lead the charge for the kind of integration that is needed to make a cloud as easy to use as your TV.