.gif)
Storage management - The software story
An efficient storage resource management (SRM) strategy can pay dividends, says Maziar Tamadon
With every new product release, the ‘building blocks’ of storage (such as storage switches, arrays and HBAs) become more intelligent and capable of better fine-grained control over how enterprise storage is configured, provisioned and secured.
But these building blocks need to be managed efficiently and cost-effectively, so that storage management costs do not wipe out the savings and agility an organisation seeks from its networked storage. Adding layer upon layer of storage management software on top of the physical storage hardware makes the environment more complex, and increases (rather than reduces) ongoing operational and capital costs.
A well-implemented storage resource management (SRM) strategy eliminates these unnecessary layers of software and provides an easier-to-manage, top-down view of the storage infrastructure. Such a strategy allows storage managers, from within a single application, to view their overall storage infrastructure and to perform detailed management of individual storage elements.
Storage resource management refers to the management, provisioning, troubleshooting and resolution of problems within the storage infrastructure (consisting of storage arrays, switches in storage networks, the storage fabric itself and the host-bus adapters, or HBAs, which link servers to the storage network.)
SRM tools commonly provide discovery of each element in the storage infrastructure; topology mapping which describes how the elements are connected; reporting of critical information about the devices (such as the version of the operating system or firmware installed on them), as well as statistics about the performance of each element; and event management which reports on specific events (selected by the storage administrator) such as the failure of a storage device.
They may also provide capacity management, which allows an administrator to balance the amount of data stored among various arrays to maximise performance, to lower cost or to achieve specified levels of reliability; performance management, which allows administrators to tune specific attributes such as bandwidth, the level of RAID protection on an array or port assignments to ensure devices operate at their maximum performance levels; security management, which includes zoning, LUN masking and the assignment of encryption levels to various devices; trend analysis of metrics such as the capacity available on storage arrays or bandwidth bottlenecks in a SAN, as well as the automation of routine tasks such as provisioning volumes as applications need more space or moving older, or less valuable, data to slower disk arrays or to tape libraries.
Wobbly software stack
In an ideal world, a single software product from a single vendor would provide all these capabilities in any level of detail an administrator needed, for any storage device or combination of devices.
In the real world, different SRM solutions have been built around specific hardware platforms, or to focus on different elements within the storage infrastructure. Many products touting themselves as ‘SRM’ only deliver some of the functions customers need, or work only with a certain vendor’s hardware. Therefore, many administrators have added ‘stacks’ of software atop their physical storage infrastructure to manage its different elements.
One example is path management, meaning the configuration and management of multiple data paths between servers and storage to increase performance or provide fault-tolerance. Another example is volume management software, which works at a layer above path management and allows administrators to create, delete and change the size of the volumes, which are available to applications and operating systems.
Virtualisation software is designed to make it easier for multiple servers and applications to access the same physical arrays by making multiple physical arrays appear as a single, ‘virtual’ storage resource. Virtualisation software may run on a server; on a storage controller; within a storage array or within a dedicated virtualisation appliance in the storage network.
SAN design and validation tools allow administrators to perform ‘what-if’ analysis of how different SAN designs, and the use of different components, will perform under actual business conditions and to identify compatibility or interoperability issues before the design is set in stone. Finally, there is software for managing elements within the infrastructure, such as the HBAs that link servers to storage networks.
Adding each of these management tools probably made sense at the time as a logical point solution to an immediate need. But over time that approach results in complex software stacks that are difficult and expensive to learn, administer and manage. Storage administrators are forced to switch from one application to another to perform even routine, but time-consuming tasks. This makes SRM far more complex and costly than it needs to be, resulting in excessively high storage management operating costs.
This approach also reduces efficiency by making it more difficult to balance ever changing storage provisioning needs. Rather than go through the burden of a time-consuming and error-prone re-provisioning of volumes across existing, unused storage when an application needs more space, the storage administrator may simply buy more physical storage for that application. This results in higher than necessary capital spending, the creation of yet more inefficiently-used storage, and more physical devices to manage over time – all of which increases storage operating costs over time.
An overly complex SRM ‘stack’ also reduces business agility by making it harder to expand, reduce or redeploy storage capacity as companies deploy new applications, create new business partnerships or find new ways to distribute, share and analyse data.
Unified approach
A far more effective approach is the use of a high-level SRM solution, which is tightly integrated with tools that provide detailed information, and configuration capabilities, about elements deep within the hardware stack. This gives managers a consistent and coherent view of their storage infrastructure from a single management interface, and combines the overall view of higher-level SRM applications with the power and flexibility of the software used to manage individual hardware elements.
One example would be integrating the software used to manage an HBA with an SRM tool that can view and manage the entire storage infrastructure. For example, Emulex’s HBAnyware 3.0 is used for managing HBAs, but integrates with leading SRM solutions. This allows a storage manager, from the same SRM environment, to both view their overall infrastructure, and to perform detailed tasks such as updating the firmware on every HBA in their environment.
The benefits of this integrated, centralised approach include lower administrative costs; increased storage efficiency and lower capital expenses; greater business agility and a reduced burden on over-stretched and limited storage administrators.
In looking for such ‘integrated’ stacks, many customers will have already chosen their overall SRM environment. If they have not, they should look for how well it supports the SRM features (such as discovery, automation and event reporting) that are most critical to them now and in the future. For the lower-level software for managing the individual storage elements, they should look for detailed diagnostics, the ability to provide centralised management, and the ease of use and reduced training requirements that result from close integration with the overall SRM environment.
Networked storage infrastructures are becoming more complex all the time, encompassing a greater variety of storage arrays, servers and network topologies. Each of the elements in this infrastructure, such as switches, HBAs and arrays are gaining more intelligence, which requires more and better management of each of those elements. Customers can try to manage their storage environments by adding layer upon layer of storage tools to their storage infrastructure, but this only adds to the complexity and cost that networked storage is supposed to reduce. With each new ‘point’ management solution that makes it harder to manage, configure and secure its storage, the enterprise becomes less efficient and less agile.
Customers should instead demand that their storage vendors deliver ‘element-level’ management tools that provide both highly detailed management of individual storage components such as HBAs, switches and arrays, and that also integrate tightly with high-level SRM tools. With a centralised, consistent ‘single pane of glass’ storage managers can reduce both their capital and operational expenses, and deliver the most business value from their networked storage infrastructures.
Emulex Corporation is exhibiting at Storage Expo 2006 at the National Hall, Olympia, London from 18 - 19 October 2006 www.storage-expo.com
Maziar Tamadon is Director, Product Marketing, Emulex Corporation www.emulex.com
Share this article with others
Post to del.icio.us
Printed from http://www.eurocomms.com/features/111366/Storage_management_-_The_software_story.html
Comment on this article
Skip to comments
We encourage users to analyse, comment on and even challenge European Communications's articles, including the one above - 'Storage management - The software story'
User reviews and comments that include profanity or personal attacks or other inappropriate comments or material will be removed from the site.
Additionally, entries that are unsigned or contain "signatures" by someone other than the actual author will be removed. We will take steps to block users who violate any of our posting standards, terms of use or privacy policies or any other policies governing this site.