Browse Prior Art Database

Enable usage of multiple disks sizes in a GRID storage

IP.com Disclosure Number: IPCOM000230611D
Publication Date: 2013-Aug-26
Document File: 3 page(s) / 75K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a new method and system that allow better use of the available disks in a storage GRID, both small and large, as long as the disks have the same form size. This novel approach allows a storage array to use disks of different sizes. The idea is to change the current goal of keeping the same size for all slices to a goal where all slices have the same free-space-size.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 49% of the total text.

Page 01 of 3

Enable usage of multiple disks sizes in a GRID storage

One of the fundamental principles of a GRID device such as storage array is to maintain symmetry across its components. The idea is to keep the GRID balanced and to that end, all components must be similar or identical and have an equal part out of the GRID total resources in order to share the workload. For example, in a storage array that can hold up to 180 disks, each disk hosts an almost identical number of slices out of the total ~16K slices. Hence, a slice can increase or decrease in size, as long as all slices share the same size. In general, the more disks there are in the system, the bigger each slice is. When the system loses some of its disks, or is simply equipped with fewer modules, the surviving or remaining disks host the fixed number of slices. As a result, each slice reduces its size accordingly. Having smaller slices translates into less free space on the system as the used partitions within the slices are kept. The reduced size is at the expense of the free space.

As bigger and denser disks become more prevalent, new storage GRIDs are configured with the growing disks. Each generation of the storage GRID is tightly tied to the disks' size at the time it was manufactured and configured. Therefore, storage clients cannot take advantage of the bigger disks and cannot use smaller disks that are already available. In fact, as the smaller disks give way to the larger, if a smaller disk has to be replaced, the practice is to ship the new, larger disk to the client but "short-stroke" it to the smaller disk size in order to keep the symmetry among all disks.

In a market where system openness is a major buying decision factor, the pronounced inability to "upgrade" a system in the future or the inability to reuse otherwise compatible disks (though of different sizes) in the system can have a negative impact on the marketability of the hosting appliance. A feature is needed that offers a client the option to use either

newer or older disks on a purchased system.

The solution is a new method and system that allow better use of the available disks in a storage GRID, both small and large, as long as the disks have the same form size. This novel approach allows a storage array to use disks of different sizes.

The idea is to change the current goal of keeping the same size for all slices to a goal where all slices have the same free-space-size. Assuming that the slices' usage is "normally" distributed (i.e., that only handful of slices are busy or full), only a few slices are hardly used and the majority of the slices are moderately used, a client can have slices of different sizes, while all slices have the same amount of free space. The lesser-used slices are much smaller in the total size than the slices that are more used. Spreading the free space among all slices allows better utilization of the total space; no hot-slice can dominate the scene, as is the current case.

...