Tiered Storage: A Beginners Guide to Information Lifecycle Management
This guide covers tiered storage, a method of data storage comprised of two or more storage media classifications that include magnetic tape drives, optical discs (OD), digital optical discs (DOD), compact discs (CD), hard disk drives (HDD), and solid-state drives (SSD). The type of storage media chosen for storing any category of data is determined based on a few considerations such as cost, data security, data availability, data recoverability, data size, among other factors.
Architecture, Methodology, and Practice
Some tech blogs erroneously use the terms “tiered storage” and “hierarchical storage management” (HSM) interchangeably. However, HSM as an architecture customarily pertains to systems that automatically transfer seldom-used data to be archived in long-term data storage units. Tiered storage, on the other hand, is more or less the core methodology of information lifecycle management (ILM). ILM as a practice is defined as the application of specific policies designed to ensure efficient information management.
The Underlining Objectives of Tiered Storage
To put it simply, the primary reason organizations invest in tiered storage is to save on storage costs. This is achieved by segmenting data based on two or more organizational values. Generally speaking, administrators consider the cost and performance of media storage against the value specific pieces of data. Files accessed regularly possess a higher-value compared to files that are not. High-value data sets are usually served through faster, higher-value tiers such as enterprise flash drives; lower-value data sets that don’t require constant access are transferred to (or written on) lower-value tiers such as tapes or the cloud.
Not Everything Is About Money
There is more to operating an organization than making and saving money. Organizations use archiving for various reasons including but not limited to governance archives (enterprise information archiving), cold storage archives (nearline storage), and active archives (storage tiering that doesn’t disrupt operations).
Many organizations have strict regulatory and compliance obligations. Solutions like content-addressable storage (CAS) and write once read many (WORM) storage devices are perfect for long-term data retention and assured authenticity.
How a Tiered Storage System Works
A tiered storage system only works as good as the number of tiers that are available to the system. For example, if there are only two storage tiers, it provides a very limited choice as to where to store certain pieces of data. If data requires a specific level of storage performance and tier two complies those requirements, it will be stored in tier two. However, if tier two doesn’t meet those requirements, the data will be stored in tier one. Over time, this will create overall performance issues for tier one as lower value data accumulates.
If a tiered storage system has three or more storage tiers, it will function more efficiently because data can be stored where it’s intended. Whether an organization requires three tiers or five, the important thing is that each tier offers performance sufficient performance requirements. This is why administrators must know how many data classes they have and what they are.
The Typical Data Classes
The most difficult aspects of building a tiered storage system are classifying data into its appropriate class, choosing the most suitable storage tier for a class of data, as well as data reclassification. One needs to remember that, over time, storage requirements for data change. Instead of merely storing data in a tiered storage system and neglecting it, one should monitor and transfer it to a storage tier consistent with its new classification. Below is a list of typical data classes most organizations have to consider when building a tiered storage system.
Mission-Critical Data | Tier 0, 1 or 2
This data class is the data required to maintain high-speed applications such as supporting transaction processing systems such as in financial trading and other time-sensitive, highly volatile environments. This sort of data must be stored in the highest level tier so there aren’t any delays in data access, which results in decreased profitability due to losing customers. Mission-critical data works best in tiers 0 (zero) and 1 (one).
For this, these tiers use ultrafast solid-state storage media configured to produce the highest performance possible. Storage efficiency is not a concern with tier 0 and tier 1 storage solutions. For organizations that have been using storage tiering long enough, they’ve already switched from tier 1 to tier 0, considered an ultra-fast, intelligent storage platform. Tiers 0 and 1 are best employed by organizations dealing in areas of business where the amount of profit made supersedes the cost of the storage.
Hot Data | Tier 3
Hot data is considered as enterprise resource planning (ERP) and customer relationship management (CRM) data. Recent emails can even be considered hot data. Though hot data isn’t as valuable as mission-critical data, it needs to be accessed frequently and without delay which means it should be stored on tier 3 media storage.
Tier 3 might not be as high-performance as tiers 0 through 2 but it offers medium to high-performance drives that are comparatively low cost to solid-state media. More organizations use hot data than mission-critical data, so tier 3 storage media is the primary choice. Tier 3 storage media offers higher capacity drives offer a lower cost per gigabyte than tier 2 but still provides a great solution.
Warm Data | Tier 4
Warm data is usually classified as data comprised of emails that are only two or three days old, recently completed transactions and other forms of data that are accessed infrequently. Basically, warm data is business information, financial data, or data used with visualization systems. Such data doesn’t need to be accessed as quickly as hot data but needs to be easily accessible. For this reason, tier 4 solutions offer high capacity SATA drives for a lower cost than tier 3 solutions using high-performance RAID arrays or SAS disks.
Cold Data | Tier 5
Cold data is usually archived and never be accessed again except to comply with legal or regulatory requirements. It could be also data that doesn’t need to be accessed but in some way or another hold some value—data for big data analysis, for example. Cold data is better suited for the lowest layer of storage media which is tier 5. With lower-cost storage media, it takes minutes or even hours to process cold data sets.
If your IT department is experiencing a data explosion, a multi-tiered storage system featuring automated data transfer is a comprehensive solution to start considering. Not every organization requires up to five tiers but most of them could use at least three and a data archive. For those choosing not to archive their data, they’re taking greater risks in the event of civil and criminal actions when producing historic files or documents is required. Contact us today to learn more about a multi-tiered storage system from RAID, Inc., which could help your organization.
