Implementing Storage Spaces Direct- Managing Data in a Hybrid Network

Storage Spaces Direct uses local- attached drives on servers to create highly available storage at a minimal cost of traditional storage devices (SAN or NAS). Storage Spaces Direct uses regular hard drives that are connected to a single node of the failover cluster, and these disks can be used as storage for the cluster.

To understand how Storage Spaces Direct truly works, I think it is better to first understand some other technology terms for Windows Server 2022. When an IT administrator takes a bunch of physical disks and puts them together, it is called a storage pool. Storage spaces are virtual disks that are created from storage pools. Storage Spaces Direct is the evolution of Storage Spaces.

Many of the same features are used in Windows Server 2022 like Failover Clustering, Cluster Shared Volumes, and SMB.

Storage Spaces Direct uses disks that are connected to one node of a failover cluster and allows for the creation of pools using those disks by Storage Spaces. Storage Spaces Direct streamlines deployment by using converged or hyper-c onverged architecture.

Virtual disks (spaces) that are constructed on a pool will have their mirrors or parity (redundant data) span across the disks using different nodes of the cluster. Since replicas of the data are spread across the disks, this allows for access to data in the event a node fails or is going down for maintenance.

You can implement Storage Spaces Direct in virtual machines, with each VM configured with two or more virtual disks connected to the VM’s SCSI Controller. Each node of the cluster running inside the virtual machine can connect to its own disks, but using Storage Spaces Direct allows all the disks to be part of the storage pool that spans the entire cluster node. For the redundant data (mirror or parity spaces) to be spread across the nodes, Storage Spaces Direct uses SMB3 as the protocol transport.

Networking Hardware To communicate between servers Storage Spaces Direct uses SMB3, including SMB Direct and SMB Multichannel over Ethernet. It is recommended that you use 10+Gbe with Remote- Direct Memory Access (RDMA), or either Internet Wide Area RDMA Protocol (iWARP) or RDMA over Converged Ethernet (RoCE).

Storage Hardware 2–16 servers with locally attached SATA, SAS, or NVMe drives.

Must have at least two solid- state drives on each server and at least four additional drives.

SATA and SAS device should be following a Host Bus Adapter (HBA) and SAS expander.

Failover Clustering To connect the servers, Windows Server 2022 uses the built- in clustering feature.

Software Storage Bus Storage Spaces Direct has a new feature called Software Storage Bus. This allows all the servers to see all of each other’s local drives by spanning the cluster and establishing a software- defined storage structure.

Storage Bus Layer Cache The Software Storage Bus joins the fastest drives available to the slower drives to provide server- side read/write caching that speeds up the I/O and boosts data.

Storage Pool The storage pool is the collection of drives that form the storage space. It is created automatically and all qualified drives are discovered and added. You should use the default settings on one pool per cluster.

Storage Spaces Storage Spaces offers fault tolerance to virtual disk using mirroring, erasure coding, or both. Think of it as distributed, software-d efined RAID using the drives in the pool. These virtual disks normally have resiliency when two synchronized drives or servers fail.

Resilient File System (ReFS) The Resilient File System (ReFS) is Microsoft’s latest filesystem and was designed to maximize data availability, efficiently scale to large datasets across varied workloads, and provide data integrity. It includes improving the VHDX file operations such as creating, expanding, checkpoint merging, and built-i n checksums to distinguish and fix bit errors. ReFS also introduced real-t ime tiers, based on usage, which will rotate data between “hot” and “cold” storage tiers.

Cluster Shared Volumes The Cluster Shared Volumes (CSV) filesystem unites all the ReFS volumes into a single namespace available through any server. This namespace allows every server and every volume to look and act like it’s mounted locally.

Scale- Out File Server This option is necessary in converged deployments only. It offers remote file access by using the SMB3 protocol to clients over the network. It essentially turns Storage Spaces Direct into network- attached storage (NAS).

To see step- by- step instructions on configuring and deploying Storage

Spaces Direct, visit Microsoft’s website at https://learn.microsoft .com/en- us/windows- server/storage/storage- spaces/ deploy- storage- spaces- direct.

The Benefits of Storage Spaces Direct

The following are just some of the benefits of using Storage Spaces Direct with Windows Server 2022:

Simplicity In less than 15 minutes, you can go from a standard server running Windows Server 2022 to creating a Storage Spaces Direct cluster. It’s just the click of a check box if you’re using System Center.

Unrivaled Performance Storage Spaces Direct exceeds 150,000 mixed 4k random IOPS per server with reliability, low latency, built-i n read/write cache, and support for NVMe drives that are mounted directly on the PCIe bus.

Fault Tolerance Constantly available built- in resiliency that will handle drives, servers, or component failures. Chassis and rack fault tolerance can also be configured for larger deployments. There are no complex management steps needed when hardware fails. Simply change it out for another one and the software will fix itself.

Resource Efficiency Greater resource efficiency with Erasure coding delivering up to 2.4× more storage. Using Local Reconstruction Codes and ReFS, real- time tiers extend to hard disk drives and mixed hot/cold workloads, all while reducing CPU usage to give the resources back to the virtual machines where they are needed.

Manageability Keep excessively active virtual machines in order by using Storage QoS Controls with minimum and maximum per- VM IOPS limits. Continuously monitor and alert by using the built- in Health Service. There are also new APIs that make it easier to collect cluster- wide performance statistics and capacity metrics.

Scalability For multiple petabytes of storage per cluster, you can increase up to 16 servers and add over 400 drives. To scale out, you just have to add drives or add more servers. Storage Spaces Direct will automatically add the new drives and begin using them.

Deployment Options

When using Windows Server 2022 and installing Storage Spaces Direct, you can choose  between two deployment options:

Converged In converged, there are separate clusters for each storage and compute. The converged deployment option, also called disaggregated, puts a Scale-O ut File Server (SoFS) on top of Storage Spaces Direct to provide NAS over SMB3 file shares. This allows for scaling computer/workloads separately from the storage cluster. This is essential when working with large- scale deployments such as Hyper-V  infrastructure- as- a- service (IaaS).

Hyper- Converged In hyper- converged, there is only one cluster for storage and compute. The hyper-c onverged deployment option runs the Hyper-V virtual machines or  SQL Server databases directly on the servers delivering the storage, storing files on the local volumes. This removes the need to configure file server access and permissions. It also reduces the hardware costs associated for small- to- medium business or remote office/branch office deployments.

Leave a Reply