What is Geo-Redundancy?
Geo-redundancy is the practice of replicating and backing up your data, applications, and servers across a number of locations in the real world. Its goal is to spread your risk across multiple or “redundant” points – avoiding any single point of failure.
Geo-redundancy, or geo-redundant storage, is the practice of distributing critical infrastructure components (such as servers) across multiple data centers in different (geographical) locations. It replicates your data to another physical location elsewhere in the world, acting as a safety net against regional outages. This way, your services will be online, even with local issues.
But while its primary purpose is to safeguard against catastrophic events like natural disasters or localized attacks, it’s about more than backup; it also aims to improve network performance, by locating resources close to where people need them, decreasing latency and load-balancing network resources. There are several levels to a georedundancy strategy – and it’s not just about keeping your systems safe. Let’s go into more depth.
What Are the Benefits of Geo-Redundancy?
Making your infrastructure geographically redundant carries multiple advantages. At the technological level, your data is replicated across several locations, meaning that if one location goes down – anything from a server fault to a natural disaster – your business information isn’t affected. (When this is well implemented, actual users won’t even notice the switch from one resource to another.)
Geo-redundancy also keeps those resources dynamic – not “overnight backups”, but on-the-fly, 24/7 resource management, making sure people maintain access to their apps and data even if several critical events happen at once. After all, the best disaster recovery strategy is one that doesn’t interrupt your business functioning at all.
Good geo-redundancy practice means redundant capacity as well as copies of data. If the servers in your Paris datacenter max out, redundant servers in Madrid can share the load. And there are benefits in keeping web applications physically close to where their users are, too. If you have 1,000 people in Miami and another 1,000 in Seattle, the best place for their data isn’t your New York HQ.
Geo-redundancy also has a role in cybersecurity. Many attack vectors are geographical, hackers targeting specific sites and centers. If you maintain backup resources in a different city or country, you can mitigate even the worst breach of your defenses.
Geo-redundancy, and its benefits, build confidence and trust in your company. It keeps your people productive, your customers satisfied, and your corporate governance compliant with data privacy laws. Protecting your business information, your profits, and your reputation too.
What Are Geo-Redundancy Strategies?
Active-Passive: Simple but Limited
Active-passive geo-redundancy is the simplest approach, and the closest to “overnight backups”. It aims to keep spare resources ready and waiting, or “passive”– whether that’s capacity in an SLA, or physical copies of applications and data.
In brief: when disaster strikes, that spare capacity can swing into action, becoming your “active” technology infrastructure. It can be a low-cost approach, since the main infrastructure your users use is just writing data to the backup until it’s needed.
While a great first step for anyone without georedundancy already, active-passive isn’t the strongest protection for your assets. Because it relies on switching all your services over to an alternative location – creating a new attack vector for bad actors. And the new geographical location of your services may not perform as well for your users. That’s why it’s one of several approaches.
Partial-Active-Active: Value in Compromise
Partial active-active geo-redundancy goes some way to solving the problems of an active-passive approach. Your apps and data are still backed up elsewhere – often in several locations – but it’s not a straight backup: the data is actively in use at all locations, reading and writing data as a team with many copies creating the redundancy.
Even this, however, isn’t perfect. Because while it maintains multiple instances of your data in different locations, this approach often has a centralized location to write that data from – creating traffic jams on the network, as the infrastructure writes the same set of data multiple times.
And if you rely on real-time data, as with financial transactions, such latency can cause problems if there’s a sudden major issue. That’s why partial active-active strategies are finding fewer and fewer use cases as time goes on.
Active-Active: Full-Featured With a Strong Business Case
The gold standard for geo-redundancy is fully active-active geo-redundancy. Across your technology infrastructure, each application and data source is active and serving users at all times; read and write traffic is balanced across the network as a whole, not just written centrally.
Short version: if there’s an issue at one location, data simply flows around it to the next available resource, and the network stays up.
In the past, this gold standard carried high costs, since configuration and setup needed critical skills. But many new cloud databases are built from the ground up with an active-active approach: deploy a dozen databases around the world as part of a single infrastructure, and they’ll maintain georedundancy automagically. Which keeps all your applications and data performing seamlessly at all times.