Tiered storage can reduce data storage costs, improve efficiency, and have the ability to reuse old devices.
Tiered storage is a storage method that involves storing data on a series of different storage media with different characteristics, such as performance, cost and capacity.
Different storage media are organized in a hierarchical structure. The storage media with the highest performance is considered tier 0 or tier 1, followed by tier 2, Tier 3, and so on.
Tier 0 or tier 1 usually consists of flash memory or 3D xpoint based solid state drives, while continuous tiered storage levels may involve high-performance fibre channel or SAS drives (or raid arrays containing them), lower performance SATA drives, optical disk, tape storage systems, and cloud based near line or offline storage systems.
Before SSD hard disk and cloud storage became commonplace, the three-tier storage model using disk and tape to provide tier 1, tier 2 and Tier 3 storage may be the most popular tier storage model.
However, it is not uncommon to use a tiered storage model with five or more tiered storage levels. Each has subtle differences to produce different combinations of three key storage attributes: cost, performance, and capacity.
Objectives of tiered storage
Regardless of cost, enterprises will use SSD hard disks to meet all their storage requirements because they provide very high performance and reliability.
However, in the real world, storage costs are very important because it departments seek to minimize costs and maximize efficiency within their budgets and organizations as a whole. Unfortunately, SSD hard disk storage is more expensive than mechanical hard disk storage and much more expensive than tape storage.
This means that SSD hard drives need to be used with caution and only for storing data used by systems that require very high performance. Less important data can be stored on lower cost and lower performance systems (such as mechanical hard disks), while rarely accessed data or data retained only for compliance purposes can be degraded and stored to very low-cost offline storage systems.
Therefore, tiered storage is not driven by the needs of IT departments for operational reasons, but purely for financial reasons. Then, the goal of a tiered storage system is to minimize storage costs by using the lowest cost storage option that provides the lowest performance required.
How does tiered storage work?
A tiered storage system with only two storage tiers provides storage managers with a very limited choice, rather than where a given data should be stored. If the data needs a certain level of storage performance, it will be stored in Tier 2 if the required performance level is met, otherwise it will be stored in more expensive tier 1.
A tiered storage system with three storage tiers may be more efficient because if Tier 3 provides the required level of performance, the same data can be stored in Tier 3. If not, it can be stored in Tier 2 (if available) to achieve the required performance level, and only when neither tier 2 nor Tier 3 can provide the required performance level, it needs to be stored in the most expensive tier 1 storage.
In fact, the more storage tiers available, the more effectively data blocks can be stored in the appropriate storage tier that meets their performance requirements, rather than providing unnecessary high-performance storage tiers at an unnecessary high cost. Using techniques such as creating raid arrays or short stroke hard disks, you can create new layers from a different set of storage media types. For example, both technologies create higher performance storage at the expense of lower storage efficiency, resulting in higher financial costs.
The biggest challenge for tiered storage systems is to classify data into multiple categories, determine the storage tier most suitable for a given data category, and periodically reclassify data as data ages.
The key point here is that the storage requirements of any given data may change over time. Therefore, as long as the data no longer needs the performance of the storage tier it currently finds, it must be monitored and moved to a lower (and lower cost) storage tier, rather than storing the data in a tiered storage system and forgetting it.
Typical data classes for tiered storage include:
• mission critical data. This kind of data always needs to be stored in the highest level of tiered storage because it needs to support high-speed applications and may support customer transactions. Delays in accessing data will lead to the loss of business or have a negative impact on profitability. Performance is critical.
• hot data. This kind of data requires a higher level of tiered storage because it is constantly used in applications such as customer relationship management (CRM), enterprise resource planning (ERP) and even e-mail, and is necessary in the daily operation of the enterprise. In this tier of tiered storage, performance is important, but cost is also a consideration.
• warm data. This class includes some older data, such as e-mail over a few days, or data of completed transactions. This kind of data is accessed relatively infrequently, but it still needs to be easy to access when needed. In this tier of tiered storage, the most important consideration is cost, but the minimum performance threshold must be followed.
• cold data. Such data may never be accessed again, but it needs to be archived and retained to meet regulatory or other legal requirements, or simply because it may have some value at an uncertain time in the future, perhaps for big data analysis. Cold data is very suitable for the lowest tier of tiered storage, where access time of minutes or hours can be accepted, and low cost is the primary consideration.
In addition to the smallest businesses, it is clear that manual storage tiering can be time-consuming, cumbersome and unsuccessful. As a result, most tiered storage systems rely on automated storage tiering using a tiered management system that monitors data throughout its lifecycle and automatically moves it to the storage tier as it cools.
The software can be provided with and run on a storage system, or it can be an independent solution. It can work normally in the storage infrastructure of the whole organization, and in some cases, it can enter the cloud through the cloud storage gateway.
Tiered storage type
What does tier 1 mean?
In some organizations that use storage tiering, new storage technologies have emerged that are faster than those used in their existing tier 1 tier. Therefore, this technology is used as “tier 0” storage. However, for the purposes of this article, it will be assumed that tier 1 represents the storage tier with the highest performance.
This storage tier is designed to store highly volatile and time sensitive data and needs to be available in the shortest possible time. It is usually used in the financial trading environment or other business areas. These businesses can produce significant losses in a fraction of a second, and only the fastest storage is enough. Therefore, tier 1 storage usually requires very fast solid-state storage hard disk media. Regardless of its overall storage efficiency, it should be configured to have as high performance as possible.
This storage layer is used to store transaction data to support high-performance applications, customer-oriented systems (such as retail applications) and other systems that cannot accept very short latency. Since the required level of performance is not as high as Tier 1, a lower cost and more storage efficient solid state disk storage system is usually used to provide tier 2 storage solutions.
The third storage tier is used to store “hot” data, such as customer relationship management (CRM) and enterprise resource planning (ERP) data, and even recent e-mail, which needs to be accessed frequently without unnecessary delay. This means that it needs to be stored on media (such as medium to high-performance hard disk drives), which are relatively low cost compared with solid-state disk storage media. In many organizations, Tier 3 has much more data than tier 2 or tier 1, so high-capacity mechanical hard disks with low storage cost per Gigabyte provide an ideal solution.
This storage tier is used for “hot” data, such as data from recently completed transactions, e-mail that has been used for several days, and other data that is not accessed frequently. It can also be used to store data that can be used in business information and data visualization systems, or to create financial data required for monthly, quarterly or annual reports. But whenever you need it, you need to access it relatively quickly. The demand for Tier 4 storage may be very large, and since cost is the main consideration, the typical solution for Tier 4 storage is high-capacity institutional hard disk storage. Tier 4 storage is more likely to use low-cost, relatively low-performance SATA hard drives than high-performance SAS hard drives or raid arrays.
The lowest tier of storage is typically used to archive “cold” data, which may never be accessed again, but still retains some value (perhaps for data mining in the future), so it is worth preserving. It is also used to store data that must be retained for compliance purposes, which only needs to be accessible in days or weeks, not minutes or seconds.
Optical media or tape storage systems are usually used in this storage layer, which provide storage at a very low cost per gigabyte, but the performance is very low. In the past few years, cloud based storage (such as Amazon’s Glacier storage service) has become more and more popular. These companies want to divert a large amount of data for low-cost storage and retrieve it relatively easily (if necessary) in a few hours.
Key benefits of tiered storage
Tiered storage offers many benefits, the most important of which are:
• reduced storage costs: by storing each data class in the lowest cost storage and providing at least the required minimum performance, enterprises can avoid paying for unwanted high performance. The reduction of storage cost is the main driver of adopting tiered storage system.
• higher storage efficiency: high performance hard disk storage systems that rely on raid or short stroke have low storage efficiency because the amount of data that can be stored on these systems is less than the total storage capacity provided (in some cases, much less). Because tiered storage systems can reduce the need for RAID and short storage (by moving data that does not need this high performance to a lower storage tier), storage efficiency can be improved.
• the ability to reuse old storage devices: storage tiering can provide new vitality for the storage system, otherwise the storage system will need to be replaced because they no longer provide sufficient performance for the use of upper layer data. Instead, they can be used to provide lower tier storage.
Responsible editor: CT