By Noemi Greyzdorf
Managing information assets has become more critical as organizations seek ways to reduce costs associated with storing information, better understand what information assets exist and where, and ensure operational uptime of the infrastructure. To achieve these goals, managers need to make timely decisions about their infrastructure that will impact where information assets reside and how they are managed.
When applications and their data resided on a single system (a server running the application and storing data on internal or direct-attached storage), knowing where information “lived” and how it was being protected was relatively simple. If the ERP system was running on server A, then the data associated with the ERP system resided on the disk drives attached to server A.
As environments evolved into distributed architectures where multiple servers and applications share storage and network resources, understanding where information lived and how it was being managed became more complex. The need for tools that monitor and report on the environment and help managers identify the location, access patterns, protection levels, and availability of any given set of information became a priority.
Storage management software
Storage management software is designed to help IT organizations establish the location of data and its main attributes and map it to underlying resources. Using storage management software can help group data sets according to business units or owners, determine consumption of storage resources by a given data set, perform cost analysis on the value of data at a given point in time and the cost of resources the data is consuming, and set up policies that manage infrastructure based on requirements.
Organizations can evaluate the need for resources against the value of the information assets to the business. As the value to the business and utilization of information assets change, organizations can respond by making knowledgeable choices about the technologies they want to use to maintain and support these assets. Storage management software can serve as a centralized repository of knowledge about data and storage assets, managing not only primary resources, but also resources associated with archiving, business continuity and replication, and data protection.
Storage resource management
The storage management software market today can be divided into three categories: heterogeneous storage management, homogeneous storage management, and SAN management.
Both homogeneous and heterogeneous storage management tools support storage resource management (SRM) functionality. SRM software interfaces with the host operating system and file system to understand storage utilization from the host or application perspective. SRM can enable managers to see the types of data being stored, who owns it, how long it has been stored, when it was last accessed, etc. This software can also map data from the application to the disk drive, helping to more accurately forecast future needs, understand resource utilization, and view data paths.
Heterogeneous storage management software supports multiple vendors’ storage platforms through APIs or via the SMI-S standard. In most cases, data is gathered through an agent deployed on the host being monitored.
Products that fall into the heterogeneous storage management category include Tek-Tools’ Profiler, EMC’s Control Center, Symantec’s Command Central, HP’s Storage Essentials, MonoSphere’s Storage Horizon, CA’s BrightStor, IBM’s TotalStorage Productivity Center (TPC), and Estorian’s TSF. Although these products fall into the heterogeneous storage management category, they do vary on level of support for heterogeneous environments and focus. For example, only CA’s BrightStor, IBM’s TPC, and Estorian’s TSF support mainframe environments.
Homogeneous storage management software is a category that includes vendors of storage arrays who have developed advanced monitoring, reporting, and management tools for their systems. NetApp’s manageability suite and Compellent’s Storage Center are examples of this segment. The management software provides many useful metrics that assist administrators in their tasks of provisioning, allocating, managing, and protecting storage. These tools don’t support other vendors’ hardware, but they have the benefit of close integration with the supported hardware systems. As a result, in addition to the traditional uses (reporting and monitoring) of storage management software, administrators can also execute on the knowledge they gather. If there is an orphaned volume, for example, the same tool that reports it can be used to reclaim it. In homogeneous environments, most common in small and medium-sized enterprises, these tools deliver significant value that is easy to calculate.
SAN management software provides visibility into the SAN and focuses on managing performance and availability. SAN management overlaps with SRM in some areas, but is distinct in that it doesn’t interface with the application or file system running on the host and is more focused on helping managers understand SAN traffic, data paths, and infrastructure component interdependencies. SAN management software can be used to monitor bandwidth utilization throughout the day, track changes made to the environment, and visualize interdependencies in order to notify administrators when an event may impact other areas of the environment.
Examples of SAN management software vendors include Onaro (acquired by NetApp) and Finisar. Both offer a unique value proposition by being vendor-agnostic and delivering ongoing monitoring of SANs.
In dynamic environments, best practices dictate the use of both SRM and SAN management tools.
Having tools to provide visibility into the storage infrastructure is a critical step in managing information assets. Once an organization understands what it has and where it lives, it is ready to set policies and monitor them to ensure their timely execution. Managers leveraging storage management software to manage their environments are then ready to execute on the information gained. Without execution on the knowledge, the value of storage management software plummets. There are a number of ways organizations can realize the value of storage management through execution on the knowledge it provides.
One of the most common objections to storage management software relates to claims made by vendors that using storage management tools will address the capacity utilization rates that are in the 30% rate, help reclaim unused, orphaned, and/or over-provisioned storage capacity, and improve overall management of their storage environments. The problem with these statements is as follows:
- There are a number of reasons why capacity utilization rates often are as low as 30% and they are not always due to mismanagement or a result of unintended consequences. Applications are taking advantage of faster processors but storage has not been able to keep up. Eight years ago, an 18GB 15,000rpm drive delivered the same number of IOPS as does a 146GB 15,000rpm drive today. If an application requires 10,000 IOPS, the same number of drives have to be allocated; as a result, more capacity is allocated even if it is not needed. An application requiring only 100GB of storage eight years ago would consume 80 18GB drives to achieve 10,000 IOPS, an over-provisioning of 1.3TB. Today, the same application would still require 80 drives, but since the smallest drive available is 146GB, the amount of unutilized storage increases exponentially to more than 11TB. This is an extreme example used to demonstrate the challenge of equating capacity utilization with performance requirements.
There is no easy way to solve this problem, and many storage managers end up with underutilized storage resources to achieve required performance. There are other ways, of course, to boost performance and increase utilization rates, including intelligent cache, solid-state drives, and storage virtualization. Another reason utilization rates are often low is due to the way businesses procure for projects. It is not uncommon for a project to have a capital budget associated with it. This budget includes the acquisition of all necessary resources for the duration of the project or the depreciation schedule for capitalized hardware. In other words, businesses are often given a sum of money to acquire what they need for three years, a common depreciation cycle. The consequence is that managers purchase what they need not just for today, but for three years from now. Since budgets are often not shared across business units, project A would not share its storage resources with project B. The result is over-provisioned, under-utilized storage. Addressing this requires modifying capital budget allocations and distributions, business process re-engineering, and depreciation and accounting changes. Centralization of storage and IT services as an in-house service provider has mitigated the effect of this reality.
- When a new storage array is being deployed, in most situations the storage administrator must first define the RAID groups, then carve out LUNs, and then present multiple LUNs in a volume to the server. The layout of the storage system becomes a critical task that will define potential utilization rates moving forward. The diagram on p. 21 depicts what a typical array may look like once it has been deployed.
As the diagram illustrates, a typical array inherently will have some space not being used. Assume now that the system has some unallocated LUNs and a new application that is being deployed needs capacity. In order for the LUNs to be provided for the application’s use, they have to be in the right location and be in the appropriate RAID group. Reclaiming unused capacity has challenges associated with it. To take advantage of available capacity, volumes might have to moved around, which would require migration of data—not always a simple task. Migration may require downtime, which is unacceptable in certain environments. Large migration efforts require a significant amount of planning and capacity for staging.
One way to address these issues is to use some form of virtualization. Many managers use volume manager software to simplify storage provisioning, allocation, and reclamation. Path virtualization, which is a virtualization layer in the fabric that aggregates multiple storage arrays into a single storage pool and allows administrators to provision storage capacity from a central location, can also simplify operations. With virtualization, LUNs from different arrays can be pooled together into a single volume, thus bypassing an individual array’s limitations.
Storage management software continues to expand its support for a variety of infrastructures and applications that are common in enterprise data centers. These include server virtualization, Fibre Channel and iSCSI SANs, NAS and direct-attached storage, archiving, databases, e-mail and messaging, and collaboration tools. The environments IT managers must support are becoming more complex, and the interdependencies of infrastructure components must align with application availability and business continuity requirements. IT organizations need tools to help them manage their resources.
Storage management software is one of those tools, but to realize its benefits, organizations need to take the following steps:
- Understand your business and the applications that support it. Identify and categorize applications by their criticality to business operations. Having this level of understanding about your business will make the data received from storage management software more relevant.
- Define technologies and processes by which storage managers can execute on the knowledge they will gain through storage management software. Having the information is not enough; the ability to act is necessary for the investment in storage management to pay for itself. This might include the acquisition of technologies that can facilitate action or identification of processes to deliver desired results.
- Set up policies to guide how storage resources are to be used. Should applications defined as critical be on Tier-1 or Tier-2 storage? Where should snapshots or replication take place? What are the performance and utilization parameters that the organization wants to achieve? These guidelines will help in leveraging the intelligence gained from storage management software.
- Evaluate storage management solutions against a clear set of objectives. Make sure that the software is aligned with the organization’s focus and is licensed in a way that the organization will be able to scale while staying flexible.
In a time of economic uncertainty, investment in storage management software can be difficult to justify, but the tools are very important to an organization’s ability to maintain a healthy infrastructure. Storage management software can shine light into a complex environment, enabling decisions to be made based on actual events and metrics, and eliminating guesswork. It is not just about reporting and monitoring; it is about creating efficiencies, establishing standards, and ensuring continuity in storage resource management processes.
Noemi Greyzdorf is a research manager at IDC (www.idc.com).