Disasters happen. You can’t always prevent them, but you can be prepared for when they hit. For users of Windows Server 2016, Storage Replica can be a savior.
Storage Replica is the new disaster recovery and prevention engine included with Windows Server 2016 Datacenter Edition. For enterprises that have data residing on different racks, buildings, compounds, cities, and countries, Storage Replica makes Windows Server 2016 absolutely reliable and future ready.
Cluster-to-cluster replication at SAN level was possible earlier, and Hyper-V is capable of replicating virtual machines from one cluster to another. However, Windows Server 2016 Storage Replica takes things to another level. Here’s what makes Windows Server 2016 Storage Replica so special.
Versatile cluster-to-cluster replication system
Windows Server 2012 came off the shelf with the Storage Spaces feature. This feature was extended and made more powerful in Windows Server 2016, and called Storage Spaces Direct.
Until this point of time, storage replication was primarily designed to manage storage replication within clusters. However, Storage Replica enables cluster-to-cluster replication with unmatched versatility. This feature is available in Windows Server 16 Datacenter Edition, and Storage Spaces is a prerequisite.
Stretch cluster – mitigating downtimes
The capability of configuring stretch clusters on low bandwidth and high-latency networks makes Windows Server 2016's Storage Replica a truly powerful data backup and disaster management system.
In scenarios where geographically distant servers need to query a central database in real time, any downtime in the central core can disrupt the entire network. This becomes a key use case for Windows Server 2016 Storage Replica. Microsoft has pushed for the stretch cluster concept with Storage Replica for Server 2016. Stretch clusters are clusters with significant server separation. Storage Replica allows cluster storage to be synchronously or asynchronously replicated across sites. This means the downtime in a corporate datacenter does not pull down the dependent servers.
By default, synchronous updates are selected. You can use the Failover Cluster Manager to change the settings.
This enables better use if multiple datacenters, as workloads can be managed across different datacenters, enabling quick data access for proximate users and apps. Additionally, this facilitates better load distribution and better computing resource utilization. If a disaster makes one datacenter dysfunctional, its workloads can be quickly and temporarily moved to another site.
Windows Server 2016 Storage Replica comes with capabilities to manage several scenarios, such as cluster-to-cluster replication, server-to-server replication, and stretch cluster. Using block-level, volume-based replication, Storage Replica offers a lot of flexibility and leverages SMB 3.1.1 to deliver functionalities such as encryption, digital signatures, RDMA, and multichannel (no, this has nothing to do with you going back and forth between Discovery Channel, ESPN, and Fox News).
Replica updates are managed using transaction logs as in databases. Depending on whether asynchronous or synchronous replication mode is selected, the log is appended and synced as appropriate.
Let’s understand more about Storage Replica features and functionalities that make it one of the most versatile, secure, and power-packed disaster-recovery systems for server admins.
Security features and disaster recovery
Storage Replica eliminates any possibilities of data loss with its synchronous replication mode. Also, block-level replication means that there are no instances of file locking leading to data loss. It leverages the industry benchmark security technologies such as AES-128-GCM data encryption, packet signing, Intel AES-NI encryption acceleration, prevention of man-in-the-middle attacks, and Kerberos AES256 for intern ode authentications.
Best technology, with ease of use
Storage Replica leverages SMB 3 technology and all its proven characteristics like iWARP, SMB direct support on RoCE, and InfiniBand RDMA network cards. Storage Replica's entire feature set is used in host-based and virtualized-guest deployments. Ease of use is at the core of Storage Replica. All you need is a single PowerShell command to create a replication partnership between servers. The Failover Cluster Manager tool plays the role of an intuitive wizard to help you deploy stretch clusters (this has nothing to do with the USS Enterprise in “Star Trek” or its weapon system).
It's possible to limit Storage Replica to individual networks based on the server used, data volume replicated, and so forth. This helps manage software bandwidth better.
Instantaneous initial sync
Storage Replica delivers faster initial sync, using the seeded initial sync functionality. When some data is already present on the target system, Storage Replica updates only the missing data, and hence uses low bandwidth, apart from reducing initial sync time, and no, sync time has nothing to do with recording music! Storage Replica also blocks checksum calculations and aggregations to reduce the sync time. Also, Storage Replica uses thin provisioning to deliver almost instantaneous initial data sync in several scenarios.
Cluster Health Service integration
Network admins have often complained of how Windows Server versions don't offer easily accessible reports on clustered storage. With Windows Server 2016, the problem is solved. Windows Server 2016 TP5 is enabled with a “Health Service” that improves the routine operations, monitoring, and maintenance activities associated with Storage Replica, clusters, and Storage Replica Direct.
With many automation options, action buttons, metrics and faults, and quarantine option, Health Center is a power-packed maintenance wizard. Furthermore, you'll need to turn the option on in TP5 for mainline SR scenarios.
A word of caution from Microsoft about special purpose replication solutions: Storage Replica is not the ideal solution in such scenarios. Active Directory, for instance, is built to offer a cutting-edge solution to global replication requirements.
Goodbye low-end disaster recovery systems
Existing file replication systems like DFS Replication could soon become potential candidates for becoming obsolete. Though these low-end disaster recovery systems work well on low bandwidth, the latency is very high. Because these systems require files to close, and need an artificial throttle to keep network congestion to a minimum, the high latency is expected.
This, however, means that the latest files in the network are not likely to be quickly replicated. Because Storage Replica works below the file level, it doesn't suffer from these congestion-management limitations, and is a suitable replacement for these systems.
State of the art
Though Microsoft Windows Server has delivered cluster-to-cluster replication feature for quite some time, the Storage Replica feature does that with a lot more ease and effectiveness than any other system. This, along with other value-adding features, makes Storage Replica a highly empowering feature for server administrators.
Have questions? Read our FAQ.