High availability has been a staple feature in datacenters for years. Keeping the infrastructure running and the lights on is imperative for companies in a world where small amounts of downtime can result in massive losses. As a company’s data becomes more valuable, it becomes even more important to extend that high availability to include their storage.
This is where storage-based replication comes in. Immediate access to data becomes crucial in case of a failure at the primary location.
Replication is an approach to high availability that stores data in multiple locations. Data is immediately replicated to those multiple sites whenever a change is made or data is added. This allows you to retrieve that data from any of the sites. These sites could be physical datacenters owned by the organization, or geographically separate cloud instances like Azure or AWS.
Since data is immediately replicated to secondary storage as it is written to primary storage, losses can be minimized to seconds, or even eliminated completely should the primary site go down. Generally, this is controlled by software, but more and more storage arrays are being built with integrated replication making it more efficient and easier to set up.
Typically, integrated replication requires storage arrays to be either connected physically, or in a storage area network or SAN. The arrays need the ability to talk to each other at all times in order to send data to the other array as soon as it is written on the first. These replication times tend to be measured in milliseconds.
HPE Alletra’s Peer Persistence
HPE Alletra’s peer persistence is a great example of storage-based replication. Two Alletra arrays that exist on the same subnet can become replication partners. With response times under 5 ms, the arrays can get any data written to one array copied over to the other almost instantaneously. Add in some extra features like load balancing between storage pools and a 99.9999% uptime, companies can be sure their data is always safe.
Another way of setting up replication is to use software. There are many vendors, and even windows server can be used. This type of replication usually involves clustering hardware together to make identical environments that mirror each other and can be used interchangeably.
How is this Different from Backups?
So how does this differ from backups? A backup is a copy of data from a certain point in time. Replication is copied immediately. Any change to primary data is immediately replicated and no restore points are created. Backups create and store restore points from multiple times allowing you to recover from past data sets if needed.
Backups are critical in the event of corrupted or lost data. So, if a hacker encrypts your data, or an employee deletes files, you can recover all that from a previous point in time. You can’t rely on replication for data recovery. This is because any encryption or deletion is immediately replicated, making your second copy useless.
Both are critical to ensuring access to a company’s data at all times. As data is becoming more and more valuable to data-driven companies, so are the technologies used to safeguard that data.
Contact Zunesis to learn more about data storage protection offerings.