by September 12, 2002 0 comments



Partitioning is the ability to sub-divide a server into smaller segments capable of running a version of an OS or a specific set of application workloads. Partitioning a large system paves the way for server consolidation, testing new versions of applications and OSs, increased hardware utilization, application isolation and increased flexibility of resource allocation. Server consolidation reduces the total cost of ownership, system-management requirements and footprint size. 

Server resources can be partitioned in several ways. The term physical partitioning is used to describe a method of dividing a server into smaller processing units. These units can be considered small servers, albeit part of the same hardware infrastructure. It is called physical partitioning because the partitions have to conform to the physical boundaries of the building blocks used to construct the server. As a result, physical partitioning generally affords less flexibility when attempting to create partitions with just the right amount of resources for the needs of the applications. 

Logical partitioning differs from physical partitioning in the way resources are grouped to form a partition. Instead of grouping by physical building blocks, logical partitioning adds more flexibility and freedom to select components from the entire pool of available system resources. This allows better granularity, which maximizes the resource usage on the system and minimizes unnecessary resource re-allocation. Logical partitioning provides greater flexibility when deploying multiple workloads on servers. 

Generally, servers are sized to meet the demands of the peak workloads that they are expected to run. However, it is common for the average utilization of resources in production systems to be far less than their peak workloads. 

A highly reliable server with sufficient processing capacity, which is capable of being partitioned, can be used to consolidate multiple such smaller servers on to one system by logically sub-dividing the one server into a number of separate smaller systems. In this way, application isolation needs can be met in a consolidated environment with the additional benefits of reduced floor space, a single point of management and easier re-distribution of resources as workloads change. With logical partitioning, you can adjust a partition that experiences a heavier workload at specific times, to take on resources from other partitions that do not require them during that peak period. Adjusting the resource allocation across partitions depending on workloads on the partition can facilitate better utilization of the server.

Many OSs provide resource-management capabilities that can be applied when the OS is running with a logical partition. Resource-management options give system administrators more control over the allocation of computational resources (such as CPU, memory and I/O) to applications. In this way, workloads can be prevented from consuming all the available resources. Also, it provides a mechanism to balance the use of the system resources optimally. By grouping applications by resource usage behavior, the workloads can be managed together to maximise the utilisation of the server. 

Logical partitioning even enables different regional workloads to be consolidated on to a single server, with different OSs, and different time and date settings. For example, workloads for operations based in New York and New Delhi can run in different logical partitions on a single server. 

The nature of system partitioning leads to significant differences between traditional systems and logical partitioning-capable systems. Many considerations must be taken for each partition to provide an environment as similar as possible to a standalone server without losing flexibility or functionality. Several system components must work together to implement and support the logical-partitioning environment. The relationship between processors, firmware and OSs requires that specific functions need to be supported by each of these components. Therefore, a logical-partitioning implementation is not solely based on software, hardware or firmware; it depends on the relation between the three components. 

A logical partition-enabled server includes not only Reliability, Availability and Serviceability (RAS) capabilities, but also adds new features. These deliver an increased emphasis on hardware recovery, software recovery and predictive degraded modes of operation permitting a high degree of concurrent and deferred maintenance. From a hardware standpoint, the RAS functionality in a logical partition-enabled machine is higher than in a standard environment, as procedures have been enhanced due to the multi-server characteristics. 

Logical partitioning provides excellent isolation from applications, the OS and most hardware elements. Scientific implementation of logical partitioning can go a long way in optimizing server performance while at the same time maintaining a tab on operational costs. 

M Ganesh, Country Manager, Enterprise Systems Group, IBM India

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.

<