Enterprises are striving toward cloud like operations, often called “The Cloud Data Center”. The goal is for IT to provide the flexibility and scalability that the modern business is demanding from data center operations. The storage infrastructure has a key role to play in this movement. It needs to provide an elastic infrastructure that can scale as rapidly as the rest of the environment and it has to be able to assure storage I/O performance on a per-VM basis.
Gridstore provides a storage solution that is purpose built for Hyper-V and Windows Server 2012. This week at Microsoft’s Worldwide Partner Conference they will make two announcements that enhance their position as a key ingredient to the Cloud Data Center. First, they are providing tight integration with Microsoft Systems Center and second they are adding a 48TB node to their product offering.
Who is Gridstore?
As Storage Switzerland covered in a recent article Gridstore is a software defined storage solution purpose built for Hyper-V environments. Unlike other software defined storage solutions it uses a combination of software based controllers running on the hypervisor and ethernet attached, dedicated, storage-class hardware. Having the storage software run on each physical Hyper-V server provides excellent per-VM control and almost unlimited performance scaling. The dedicated storage hardware provides consistent performance that can also scale with the environment. Most importantly, these two components can be scaled independently, which is important since most environments don’t need to scale performance and capacity at the same pace.
Integration With Microsoft Systems Center
Providing per-VM control is critical to the success of the cloud infrastructure, and integrating with Microsoft Systems Center makes it even easier for the Hyper-V administrators to apply that control. As the screenshot below depicts, Gridstore, via System Center VMM, can now provide information and control over vPools, vLUNs and VM IOPS. Operations can be more nimble as provisioning of capacity and performance can be automated via policy or self-service. Finally, the Gridstore environment can be monitored both in real time as well as providing historical reports. Any issues that arise can now be investigated and resolved.
The hardware part of the Gridstore solution is a scale-out grid architecture based on nodes. These nodes come in two forms: performance nodes for Hyper-V, database and other demanding I/O applications, and capacity nodes for D2D backup, archive and streaming I/O. Gridstore has added a new node type for each category. Both of the new nodes are 2U instead of 1U and have a 48TB of hard disk capacity per node instead of the previous 4TB or 12TB. The performance option adds 1.1TB of flash and upgrades the network connections to dual 10GbE instead of dual 1GbE.
These new nodes should appeal to larger organizations with higher capacity requirements that are looking to eliminate “node sprawl”.
Storage Swiss Take
Gridstore seems to have found its groove by presenting itself as storage, purpose built for Hyper-V. Their solution provides VM level I/O control so that mission critical applications can be virtualized with assurance that their performance will not be impacted by other VMs. They also provide decoupled scaling of capacity and performance. Now, with tight Microsoft Systems Center integration and higher capacity nodes, they can play a key role in helping Microsoft make Hyper-V the hypervisor of choice for the Cloud Data Center.
Gridstore is not a client of Storage Switzerland