Remember Tetris? As the game pieces fell faster and faster, you juggled them frantically until... ?Game Over.?
Managing your storage environment can resemble a massive game of Tetris, but the consequences can be more painful. It?s a dilemma for every IT executive: while blocks of data continue to flood into the enterprise network, the budget of IT for storage investment remains the same, if you?re lucky. If not, it might well shrink.
?We have a yearly data growth rate of 30 percent,? said Raymond Ngai, manager, IT infrastructure at Hong Kong Jockey Club (HKJC). ?But we can?t keep increasing investment (in storage).?
HKJC shared their experience recently at Computerworld Hong Kong?s Storage Forum on how to better utilize existing storage space to store more with less.
Start with self-assessment
To ensure zero loss of data, the HKJC always allocates enough storage space for each new application. But without guidelines to prioritize data, the amount of data within the enterprise network amasses. Aiming to resolve the situation, the HKJC formed a team early this year to discuss different storage problems. With the help of a consultant, the team conducted a storage assessment exercise to look at five major applications, examine existing infrastructure and classify all data involved.
The assessment was valuable as it revealed three major problems in the organization?s storage infrastructure, said Ngai. First, with a majority of data stored using direct attached storage (DAS) equipment, they had a rigid infrastructure for sharing storage resources. The second problem was the utilization rate: a low 50 percent to 60 percent. The assessment also revealed the problem of inherited service level agreements (SLAs), in which the storage requirement has not been renewed for years.
For example, to ensure that transactional data is being stored properly, the data is first mirrored locally and sent to the disaster recovery site, where data is mirrored for the second time, making four copies for every piece of transactional data.
?We started to wonder whether it is necessary to have four copies of all our data,? said Ngai.
With the understanding of the problem, the HKJC was able to design a solution with a new set of storage policies and a tiered storage infrastructure. Under the new policy, data is being re-evaluated and allocated to different tiers according to their value, said Ngai.
The tiered infrastructure is divided into four levels, consisting of three storage area networks (SANs) and one DAS, each has different levels of storage requirements. The premium SAN has the most storage space for duplication and the highest performance of I/O requirement for more real-time mirroring.
Although there are challenges to convince users of the value of the new SLA model and storage policy, the revamp of the storage infrastructure is expected to eliminate the duplication of data. The HKJC is currently implementing the plan and aiming to increase its utilization to 70 percent.
HKJC is not the only enterprise recognizing the benefit of a comprehensive storage strategy. A survey at the event shows 38.8 percent of the participants identified improving the efficiency of storage through better management and allocation of resources as the key goal for a comprehensive storage strategy. It was followed by reducing IT infrastructure complexity (26.5 percent) and ensuring zero loss of critical data (22.4 percent).
Set your priorities right
The HKJC?s storage assessment exercise is a perfect example of realizing your data?s value importance, noted Bob Mahoney, IBM business line executive of TotalStorage Solutions.
?The idea of data value becomes particularly important when storage space is used up,? he said, ?and data needs to be (selected) for deletion.?
Realizing the value of data is significant not only in deciding what to delete, but also how much to spend. The forum survey showed that 61.5 percent of the respondents chose business value of data as the key factor in their decision to make additional storage investment.
Nevertheless, understanding your data value is just the first step in devising an effective plan to increase storage utilization. More importantly, enterprises should design their storage environment to fit different levels of data, which is what information life cycle management (ILM) is all about.
?ILM is the process of managing information,? said Mahoney. ?In a manner that aligns costs with the changing value of information.?
Since not all data is created equal and the value of data changes over time, ILM allow users to spend their storage investment accordingly. Although the concept was first introduced by IBM as hierarchal storage management in the 1950s, only in recent years has it become more popular in recognition and implementation among enterprises, he said.
With the pressure of data growth, limited budget and changing business environment, ILM provides the flexibility to manage the storage environment.
?We need to adapt to flexibility in the business environment,? said Mahoney. ?We have to bend, instead of break.?
Despite its benefits, ILM has yet to receive strong endorsement among Hong Kong users. The forum survey showed only 14.9 percent of the participants indicated ILM as the most strategic storage infrastructure solution in the company, behind other major storage initiatives like business continuance (21.3) and simplifying storage infrastructure (22.3).
SAM -- storage area management
While ILM offers better storage utilization benefits, the actual allocation of resources could be complicated, noted Vivekanand Venugopal, sales director, storage management solutions at Hitachi Data System (HDS).
Since business value changes from the data?s creation to deletion, to allocate appropriate storage resource accordingly requires a lot of effort, Venugopal explained. ?Can you imagine mapping every piece of data from different applications to different ports at the storage infrastructure?? he said.
To provision storage in a SAN environment is a long process and requires 20 steps from four different layers in the entire storage infrastructure, said Venugopal, quoting a report from Gartner.
The Storage Networking Industry Association -- a consortium of storage, networking and operating companies -- also indicated in a report that mapping storage asset with the application layer is one of the top five pains for users in storage management.
?To reduce that pain, you will need to automate that process,? Venugopal said. This is what SAM -- storage area management -- aims to do.
With the purpose of better aligning IT with business objectives, SAM is a set of applications and tools that sit across the entire storage infrastructure, from the application to the logical storage array layer, to allocate storage resources according to data value.
?Gartner says SAM is growing in importance, particularly given its potential to significantly lower the total cost of ownership of storage, which hovers near 50 percent of the total IT budget in some organizations,? said HDS on their Web site.
When selecting a SAM, most respondent from the forum indicated functionality (37.6 percent) and price (35.3 percent) as the top priority.
The network foundation
Apart from ILM, users could better utilize its storage space starting with small steps into the networked storage arena, noted Don Hansen, IBM Worldwide business unit at McData, a switching technology provider that supports the development of SAN.
Like other networked storage environments, SAN helps consolidate storage resources and centralized management of data. In addition, SAN brings value-added services like monitoring of storage environment, noted Hansen.
?Before things happen, you will receive an indication,? he said. ?It will also warn you even if you?re out of the office.?
With networked storage, a company could achieve a significant utilization of storage space of up to 80 percent and reduction of storage downtime by 30 percent, said Hansen, quoting data from The Maya Group, who conducted a customer benchmarking study sponsored by McData in January 2004.
The benefits of a networked storage scheme are being recognized by local users -- 41.5 percent of the audience indicated a need for networked storage and 46.2 percent see that need in the future.
Yet, the actual implementation is not as well received. The same survey indicated 52.4 percent still uses DAS. Only 33.3 percent said their company has a SAN or network attached storage (NAS). The number of implementations is however expected to increase, as 14.3 percent said the company is currently moving to a networked storage environment.
Don?t forget the SMEs
ILM and networked storage plans may not be adequate for most SMEs, said Wilson Chin, product marketing director of Asia Pacific and Japan for EMC, as not many hardware offerings cater to smaller scale operations.
With the focus on enterprise markets, many (47.8 percent) of the event?s participants indicated that the current storage offerings do not address Hong Kong?s SMEs requirements.
While most SMEs face the same challenges as enterprises in storage management -- like data recovery, improvement of system reliability, back up efficiency and 24x7 availability -- their requirements are different.
?With a lack of IT resources,? Chin said, ?ease of installation and maintenance outweigh the importance of the feature-set in a SME environment.?
To better utilize storage resources, he said, a storage consolidation driver with pre-loaded software -- which can provide disaster recovery, DAS, SAN and NAS connectivity in a size of a blade server -- is the best fit for SME users.
Apart from better utilizing your storage infrastructure, another way to address the central dilemma is better control of your IT spending.
With an increasing demand for IT service quality but a shrinking IT budget, a charge-back system for business users is another way to achieve better utilization of storage, said Cheng Ming Daw, director of consultant and education at Veritas.
Under the charge-back system, users will be charged according to the amount of storage resource they require, instinctively aligning IT with business needs.
The first stage involves dedicated systems infrastructure, where no SLA is used to govern IT service quality. By creating shared storage, companies have reached the second stage with a centralized software management with some form of SLAs in placed. It is followed by the assisted management stage, when a shared server pool is developed with basic auto provisioning available.
As companies implemented ILM, a hierarchical modular architecture, they have entered the fourth maturity level, where a charge-back mechanism is available. When IT is fully automated to provide services according to changing business needs, companies have reached an appropriate maturity level.
Chen quoted Sky Television in the UK as the example when it developed different classes of services in its disaster recovery strategy. The company developed its services into five different classes, from platinum level, where critical data and application requires immediate recovery, to copper level for gradual recovery. Different SLAs, technology, storage media is then assigned into each level accordingly.
The result was rewarding, said Cheng, as the company is able to provide a clear picture of how disaster recovery cost is distributed among business service.