VMware Software-Defined Storage. Martin Hosken. Читать онлайн. Newlib. NEWLIB.NET

Автор: Martin Hosken
Издательство: John Wiley & Sons Limited
Серия:
Жанр произведения: Зарубежная образовательная литература
Год издания: 0
isbn: 9781119292784
Скачать книгу
for Storage Resources

As illustrated in Figure 1.6, the operational management, disaster recovery, and environmental costs are the real drivers behind the total cost of ownership calculations for storage devices.

Figure 1.6 Breakdown of total cost of ownership of storage hardware

      One of the factors that contributes to these operational costs is the heterogeneity of enterprise storage infrastructure. This significantly increases the challenges associated with providing a unified management approach, and as such, increases costs. Some IT organizations use this as a driver for the replacement of their heterogeneous storage platform, in favor of a more homogeneous approach. But typically, the replacement environment results from a deliberate attempt to procure the latest, best-of-breed technologies, or an attempt to facilitate storage tiering through a combination of hardware from a variety of vendors. Storage vendors often are unable to offer a varying portfolio of products for different types of workloads and data use cases. Furthermore, this problem is exacerbated by vendors who offer a wide range of products but can’t typically offer a common management platform across all storage offerings. This is especially true when vendors have acquired the technology through a business acquisition.

The simplified formula in Figure 1.7 can be used to estimate the annual total cost of ownership of storage resources over the hardware’s life expectancy.

Figure 1.7 Simplifi ed annual total cost of ownership

      The next aspect of calculating storage costs relates to how efficiently storage capacity is allocated to the appropriate storage tier. Utilization efficiency, provides a measure of how effectively storage capacity is allocated to the correct storage type, based on factors such as frequency of access, availability requirements of the data, or required response time.

      IT organizations often do not use storage capacity efficiently. They often use tier-1 storage to host data for workloads and applications that do not require the expensive, high-performance attributes that the hardware is capable of delivering. Tiered storage is supposed to enable the placement of data based on cost-appropriate requirements for performance and capacity, as defined by the business. However, a growing movement to flatten storage via strategies such as those offered up by Hadoop (among others) in leveraging the hyper-converged storage model is, in itself, eliminating the requirement for tiered storage altogether.

      IT organizations are typically charged with identifying the tiers of storage, the storage technologies employed, and the optimal percentage of business data that should represent each category of storage. Failing to do so undoubtedly leads to a significant increase in the per gigabyte cost of storage, which in turn results in an inflated total cost of ownership for the storage platform.

Figure 1.8 shows a tiered storage example. A business’s IT organization uses this cost per gigabyte model to determine cost-appropriate storage for a specific workload type. For instance, if the IT organization requires a 100 TB storage estate, employing only two tiers of storage (tier 1 and tier 2 in this example), the total disk cost would be approximately $765,000. However, meeting the same storage requirements through the four tiers shown, segregated using the ratios illustrated, would cost approximately $482,250, and therefore represent a savings of $282,750, or a 37 percent reduction to the original cost.1

Figure 1.8 Storage cost per gigabyte example

      As you can see, an enterprise IT organization that fails to use this type of tiered storage strategy – which moves data across the storage estate based on its access frequency and other criteria – will suffer from poor utilization efficiency, as well as a significantly increased total cost of ownership of storage resources, within their storage platform.

      Information Lifecycle Management

      Information Lifecycle Management (ILM) is the primary approach used by businesses to ensure availability, capacity, and performance of data throughout its existence. When designing a storage solution for business systems, one of the key business requirements that you must understand is the ILM strategy being used by the customer for their business data.

      Modern businesses and organizations must address the challenges associated with information management and its ever increasing growth, because business data and the way that it is used is playing a growing role in determining business success. For instance, companies such as Amazon and Rakuten are using their business data to gain strategic advantage over their competitors. The use of customer profiling and identifying what a customer may wish to purchase, based on their purchase history, provides a serious competitive advantage. In addition, understanding each customer’s purchase habits (such as typically making all orders within the same few days each month, after payday) enables these businesses to target specific products at specific customers at a precise time, via a customized email based on the individual’s purchase history and purchasing profile.

Another key consideration is how the value of the data changes over time. For instance, if a customer stops making purchases or closes their account, legislation might require that data to be deleted after a set period. Therefore, information that is stored may have a different value to the business, depending on its age. Understanding how an organization uses its data, and the value of its information throughout its life cycle, can be at the heart of storage design for many businesses (see Figure 1.9).

Figure 1.9 Information Lifecycle Management key challenges

      It is also important to recognize that ILM is a strategy adopted by a business or organization, and not a product or service. This strategy must be proactive and dynamic, in order to help plan for storage system growth, and also must reflect the value of the information to the business.

      Implementing an ILM strategy throughout a large organization can take a significant period of time, but can deliver key benefits that directly address business challenges and information management and utilization. The key design considerations that relate ILM strategy to the architecture of a storage platform include the following:

      • Improving utilization by employing tiered storage platforms, and providing increased visibility into all enterprise information, alongside archiving capabilities

      • Providing simplified storage management tools and increasing the use of automation for daily storage operational processes

      • Implementing a wide range of backup, data protection, and recovery options to balance the need for business continuity with the cost of losing data

      • Simplifying compliance and regulatory requirements by providing control over data placement, and knowing what data needs to be secured and for how long

      • Lowering the total cost of ownership while continuing to meet the required service levels demanded by the business, and aligning the storage management costs with the value of the data, so that storage resources are not wasted, and unnecessarily complex environments are not introduced

      Providing a tiered storage solution that ensures that low-value data is not stored at the same cost per gigabyte as high-value data

      Implementing a Software-Defined Storage Strategy

      As a consequence of the ever-increasing cost of enterprise business storage, as outlined previously, more IT industry attention than ever before is focused on new storage architectures and technologies