Understand Network Storage Tiering
The storage process in a VDI environment is entirely different from how it is in a PC environment. While the virtual machines can be created in minutes and user interface gets active even faster, there are certain rules governing the storage. There is a concept of Network Attached Storage (NAS) decides being leveraged by many businesses today. It enables the administrators to make use of the network storage, while they cache the very crucial and urgent data on their devices, ahead of the schedule. For NAS devices, the write I/O load is considered more prior compared to the Read I/O load. This means that as the user number increases beyond a certain limit and the shared resources reach approximately 40% of the utilization, the Write I/O activities are queued in front. This degrades the Read I/O performance and the user experience as well.
To tackle this, there is a solution coming up in the form of caching strategies. Employing a front end caching solution that can have the Read I/O activities cached on to a local storage magnifies the performance by a multitude. There are companies currently working on this regard, with a hopeful indication of a successful storage caching for VDI environments.
Benefits Of Network Storage Tiering
There are two great benefits that can be immediately understood. One is about having a low cost solution which can store Write I/O workloads effectively. This dramatically reduces the storage cost when the data size is considered on an organizational level. For smaller enterprises, the benefits are lesser realized, since the storage proves to be enough for the amount of data being dealt with, everyday.
Another benefit is in terms of using the inactive data sets efficiently. The Read I/O data sets which are inactive, and yet consume a good amount of space, can be moved on to cache storage for the performance enhancement. Depending upon how critical these data sets are, the administrators and the storage managers may opt for further sage of the cached component.
Space Saving Technologies In A VDI Environment
Considering the wide span of usage that is behind deployment of VDI, any technique that can save the storage space consumption is highly welcome. Currently, there are two such techniques that fit well for VDI deployments:
1. Thin Provisioning:
This technology concentrates on the total volume of data that can be held by the host server, irrespective of the data allocation done for the client machine. For example, if the client machine is a desktop allocated with 30 GB of data, and the host writes 15 GB of data on the disk, the total allocation that appears to the desktop is limited to 15 GB. In a VDI environment, this is an efficient way to keep the resources under the specified limit of usage, while exhausting them to the fullest extent. This reduces any kind of resource wastage, and lets the client machine experience an uninterrupted performance with the controlled volume of data transfer.
The base of this technology goes back to the time when desktops were being improvised. Usually, the desktops in an enterprise were created using a master image. This image when replicated to create multiple desktops suffered with resources duplication, which was nothing more than the unneeded consumption of space. To deal with this, the de-duplication strategy was deployed that made use of linked clones and images. All of the desktops then shared the common files through the linked clones. In a VDI environment, the same technique works well to clone the storage array. Since, most of the content is the static operating system data, de-duplication through cloning is an excellent way of reducing space.
What Is Good- SAN or NAS?
Let us look at the basic difference. The SAN disk stores data at the block level. NAS stores data at the file level. Both of these storage schemes are good for VDI, though, the implementations differ. In block based solutions (SAN), the VDI layer takes care of cloning and de-duplication, while in a file system (NAS), this is performed by the storage array. Considering this distinction, the decision may be taken on where the user wants the task to be delegated.