The deployment and solution of cloud storage automatic layered system

Automated layered Systems (automatedtieringsystem,ats) migrate data between different tiers of storage. If the data is dynamic, it is migrated to the upper tier of storage and is eventually stored in a solid state disk (SSD). There are many types of automated layered systems, with the least impact and the safest way to use them as a cache for storing dynamic data.


　　

The
cache type automatic layered system copies Dynamic data from traditional mechanical storage to a cache (RAM or Flash state disk) based on high speed memory. In this copy mode, an automated layered system is used as a large read cache, with little or no unique copy of the data. Even when they are writing to the accelerator by caching inbound writes, it only takes a few minutes to keep a unique copy of the data. In these modes, these systems can help cloud storage technology serve more mainstream storage requirements.


　　

The
cloud storage system is also divided into several forms. One of the more common is the "value layer" of NAS storage, which is highly cost-effective and highly scalable. But this cost-benefit and scalability is often at the expense of performance, making cloud-based storage systems unusable for more mainstream storage resources. Many users and providers want to be able to deploy cloud storage more widely and use automated tiered systems to fill this gap.


　　


However, when used for more mainstream purposes, cloud storage systems pose a challenge, that is, they are usually software solutions, sometimes using common hardware and disk drives that are available to users. This lowers costs because the dataset is distributed across multiple types of storage hardware. Major storage manufacturers are now focused on delivering automated tiered systems to their tier-I storage platforms, while controlling costs while improving performance. However, primary storage is not commonly used in cloud storage deployments, and major storage manufacturers insist on providing automated tiered system technology in their low-end storage systems to prevent these performance upgrades from affecting their tier one storage market.


　　


balance performance gap and heterogeneous storage requirements a set of automated tiered system solutions for storage providers dedicated to delivering automated tiered systems on an external storage device. You can then place this solution at the front of the cloud storage network, but you may have to face some of the challenges of user or application access to the data.


　　


Once the automated tiered system solution is deployed, all network transmissions will pass through the system. Automated tiered system devices analyze storage transmissions and, depending on their access characteristics, keep dynamic blocks of data in a high-speed storage layer-usually RAM or SSD, or high-speed SAS. Because the read operation of these data comes from the high speed storage area, it can be quickly paid to users or applications.


　　


in NAS-based cloud storage, there are many instances of application that require higher performance. The first is to install a NAS cloud storage system that is lower cost and more scalable than traditional NAS, and an internal private cloud storage system. In this instance, it is inevitable that more performance than the cloud storage system itself is designed to deliver. Cloud storage Front-End to add automated tiered systems can often solve most performance problems.


　　


The second use instance is the more classic cloud storage provider model. If some users of a provider have some data that suddenly becomes very dynamic, the data can be migrated to an automated layered system. While most requests for this data may be accessing the data from a lower-speed connection, the 1000-user access combination can lead to storage bottlenecks.


　　


Most cloud storage systems are "loosely clustered", which means that the performance of a single node can be a bottleneck because the data is not distributed to nodes like a tightly paired cluster. As a result, if a file is accessed frequently, it can only be read from one node at a time. The solution is to copy the file to multiple nodes in the cluster and then change the application to see who else needs the file. In addition, if the frequency of access to this file is reduced, a redundant copy of the file needs to be found and deleted. In most cases, the last step rarely occurs, which leads to a lot of wasted space. This requires additional management time for the storage administrator.


　　


another simpler and more efficient solution is to add an automated tiered system. The system tiered system migrates frequently accessed files (or file fragments) to ram or to a cache based on a solid-state disk. Then, when the file is accessed frequently, the system supplies the file from the cache. This approach does not require changes to the environment (or a limited number of changes) that can be identified and migrated to the high speed store when the file is accessed frequently. Then, as the frequency of access decreases, the files are automatically migrated to the cache. As a result, storage becomes manageable and autonomously regulated.


　　


Automated tiered system solutions are often used to accelerate high-end NAS. These systems already have a high-speed disk subsystem and multiple high-speed network connections. High-end NAS is used to deliver the best performance that a mechanical drive can provide. Automated tiered systems are often used as a last resort before replacing an entire storage array.


　　


on the other hand, NAS cloud storage systems do not necessarily have the same level of performance as traditional NAS. As noted earlier, the focus is often on cost reduction and scalability at the expense of performance. As the cloud storage environment expands-or as cloud storage is more used in mainstream applications, the lack of raw storage performance forces storage managers to consider choosing more traditional solutions. One option they can consider is to not limit the automatic layering of storage applications. The combination of these two technologies will provide higher performance while maintaining cost and scalability advantages.