Managing system and information resources is a challenge for infrastructure managers. The need to ensure that users can access their systems and data means that service levels must be high and access always available. But the content of the files and databases must be managed by the businesses themselves; they understand the characteristics and the corporate, regulatory and legal importance of the information stored.
There is a dichotomy here. Sharing responsibilities often means that policies for managing information content are not clearly set, resulting in more and more data resident on expensive resources. The resultant impact is that when data protection practices are performed, much more data is being copied and kept than is necessary. And when it comes to recovering from a failure, the recovery procedures can become more protracted than is desirable. This perpetual circle results in more expensive procedures being used to keep all data live.
The first phase of consolidation concentrated on rationalising the number of arrays and servers. Harnessing the benefits of storage networking, servers and storage arrays did not have to be oversized to handle peak loads, resulting in improved utilisation of the system resources.
As organisations are going through the development cycle, we can identify a second wave of consolidation taking place, which is still realising some gains in reducing the number of system components. The key characteristic of this phase is the introduction of management tools and facilities to enable even greater utilisation of resources and to establish automated processes that will support the ever-increasing sophistication of systems.
Management policies must be established
The development of automated facilities such as performance monitors with automated reconfiguration options, data movement or archiving based on retention or usage, resource management with thin provisioning options and journaling to recover to a known state, all require clearly-understood policies as to how they will operate within each and every organisation.
Virtualisation is also enabling organisations to utilise resources differently. For example, tape virtualisation enables increasing levels of data to be stored on disk for rapid recovery. If the number of data images is not controlled, then the need for disk cache will escalate.
In the same way that security is addressed with a set of access control policies and procedures, organisations will need to establish policies to enable the automated tools to move data to a lower-cost media. If data has not been accessed for a given period, operational policies must be defined for archiving data and how it is accessed and retrieved.
These are examples where tools can assist in automating processes. In the same way that an employee needs to have a clear set of guidelines to address his/her responsibilities, infrastructure managers and business unit managers must establish operational and business rules for managing system resources, availability and access to the information itself.
Facing the challenges of the second phase of consolidation
There is a continuing rationalisation and consolidation of system resources to meet the business imperatives to control IT costs. Disk utilisation has increased through the first stage of consolidation from low figures of 10-30 percent up to 50-60 percent. But there are still additional gains to be realised.
Virtualisation is not only assisting in the better utilisation of storage resources. Server virtualisation is enabling better utilisation of processing capabilities. With the use of resource management tools, all aspects of the system infrastructure can be addressed. Service levels that measure system availability and response time do not need to be compromised when the activities and utilisation of the infrastructure are monitored and measured.
Linking this to business processes is a challenge that must be faced. The increasing volume of information that needs to be stored for the long term must be protected and secured. New tools that enable automated processes must reflect these operational needs.
The policies by which an organisation is going to operate must be clearly defined. Fixing a problem with a single remedy may work, but usually the appropriate medicine must be administered case by case.