Today, more economical flash memory promises to break the performance of storage devices in the foreseeable future.
In order to make full use of flash memory, people need to implement it in the right way and with the right technology.
In this way, the highest performance and higher efficiency can be extracted from the overall solid-state storage deployment and storage network.
For example, for active data, flash memory can provide better performance, moving parts less than hard drives.
The result is that flash memory deployment is usually cheaper than hard drives for major data use cases, especially for long-term use.
The problem with solid-state memory is that only about 5 to 10 percent of the data center data is active.
As a result, users can save some money and store the remaining 90% or more on larger capacity, less cost hard drives, or increasingly in the cloud.
Flash memory does not necessarily improve data storage efficiency and performance alone.
Users need to start with a solid foundation. Here are seven tips to improve the efficiency of data storage to make storage faster and more efficient.
(1) Improve the storage network Latency in disk-based systems does not expose network weaknesses, while indeed, flash-based systems are such, first maximizing the performance of the storage network before upgrading to flash memory or adding other SSD drives to an existing system.
The network has three components: server, host bus adapter (HBA) or network interface card (NIC in the storage system), network switch and cabling infrastructure.
It is easy to see the bandwidth capacity of the first two components (NIC / HBA and switches), which should deploy at least 10 Gbps or 16 Gbps Fibre Channel (FC) or faster channels.
While bandwidth is important, delay and delivery quality are especially so.
Most data centers do not generate enough continuous transactions to flood the high-speed network.
Instead, they have generated millions of small businesses.
The network transfers these transactions from the server to memory and returns again, which is critical for extracting maximum performance in flash investments.
(2) Data storage efficiency and performance In storage, efficiency and performance are the opposite forces, because efficiency usually adds value at the expense of performance.
Many of the technologies that people use to improve data storage efficiency (e. g., thin configuration, deduplication, and compression) actually damage storage system performance.
Flash storage creates a middle ground between efficiency and performance.
Yes, using these technologies damages performance on flash memory, just as they do on hard drives.
But because flash performance is so high, it usually provides too many performance cycles.
Therefore, from the user’s perspective, running the usual data storage efficiency program does not significantly affect the performance.
Once users have fine-tuned their storage network, they should consider using flash storage for deployment.
(3) Implement the server-side flash memory In the server-side flash design, the network and storage settings connected to the network remain unchanged, and the hard drive-based storage array is basically installed, where the speed and quality of the storage network is not as important when shared flash arrays are shared.
In contrast, server-side flash technology aggregates internal flash memory from multiple servers to create a virtual flash pool.
These server-side flash aggregation products are suitable for read and write cache, even to the storage layer.
However, they introduce network factors in terms of performance, because aggregation requires a network to create virtual storage pools.
(4) Deploy the network cache Unlike storage system upgrades (improving performance on a single system), network cache improves the performance of each storage system on the network.
These devices are basically located between the storage system and the server, caching the most active data.
Many network caches are available in high availability configurations, making them suitable for cache read and write I / O. Users can also resize the network cache so that the flash storage area is large enough to store the entire organization’s active data set, essentially converting existing arrays to archival and data-protected storage systems.
(5) Implement software-defined storage (SDS) using small flash arrays Another option to improve storage performance and data storage efficiency is to use software-defined storage (SDS).
These products run on a device or hypervisor and provide a common set of storage software features on a variety of hardware arrays.
Some software-defined storage (SDS) systems can leverage the existing storage hardware and provide an automatic migration of data between them.
If you add small flash arrays to an existing infrastructure, you can use SDS to automatically move the most active datasets to the array to improve performance, and simplify management as an additional benefit, as all storage management subsequently becomes unified.
(6) Optimize the application Before implementing new or enhanced existing storage systems, carefully check the application to run.
Many storage professionals find this daunting because they neither own the application nor understand the code around it.
The good news is that with some available programs, you can check the application code, provide high-quality analysis, and make specific recommendations on what and where to change.
While it can skip this step, it leads to more hardware problems.
Code-related performance issues may be masked by high-performance storage, but it will not allow flash memory to reach its full potential, forcing administrators to look for other potential performance losses, such as storage networks.
Before implementing the flash memory repair code, you can even avoid the need for flash memory first, or reduce the need to buy flash memory.
(7) Buy new full-flash memory or hybrid arrays This is ideal for deploying data centers with existing hard disk-based systems (still useful life, under original warranty), so users can redeploy these old mechanical hard disk systems and expand with new flash arrays.
However, at some times, users need to buy some new storage systems.
This means choosing between all-flash memory or hybrid arrays.
The initial decision is relatively simple: if the organization can provide an all-flash array that meets its capacity requirements (assuming the performance requirements), don’t hesitate to purchase.
To sum up, the path to improved storage performance does not begin with all-flash memory investments.
It begins with a careful inspection of the entire storage network.
Once complete, consider many other storage performance and data storage efficiency enhancement options, including certain types of flash storage deployments.
Which products are best used to work in data centers, and some IT departments may not even need to upgrade their storage systems.