These days, every smart organization is always trying to increase productivity and improve user experiences by accelerating their computer networks and ensuring their apps—and the data centers that power them—are highly available, fault tolerant, and easy to scale.
But in an age where more and more companies are spread out across the world—and many of them are increasingly turning to hybrid cloud environments—how exactly can those goals be achieved?
An Active Everywhere Scenario
Imagine you’re a social media startup catering to U.S. and European customers. An easy way to increase speed and availability would be to host all of your data and systems in both US and European data centers.
Unfortunately, this introduces several new difficulties. In the age of the General Data Protection Regulation (GDPR), for example, hosting data that pertains to European customers in the United States could create serious problems.
Thanks to the rise of hybrid cloud architectures and remote teams, however, companies can’t afford not to run multiple data centers if they wish to remain both relevant and functional.
So how do you bridge the gap?
Historically, companies had two options:
- Active-Passive. This approach involves two data centers in different regions that are connected by a network. One of these data centers is “active,” meaning it can handle read and write requests, while the other is “passive,” meaning it can only handle read requests. While an active-passive approach ensures data is accessible in different geographies, it introduces some other problems. For example, users that are far away from the active data center will experience latency when they try to perform write operations.
- Active-Active. This approach involves two data centers in different regions that are both able to perform read and write requests. In this scenario, data is replicated in both directions. While an active-active approach may reduce latency, it introduces problems of its own. Imagine a user in the United States and a user in Europe perform the same write operation at the same time (e.g., registering the same username). It’ll work initially. But when the data centers replicate, there will be a conflict.
Due to the limitations of traditional relational database management systems, a new kind of database has emerged in recent years: the distributed, scale-out database.
Distributed Databases – The New “Active”
Distributed databases allow you to take what we call an Active Everywhere approach to data management.
In this approach data is replicated across every data center you have, across geographies, whenever any read or write operation is performed thanks to its masterless architecture.
In addition to this resiliency, an Active Everywhere database also delivers:
- Linear scalability. Add additional nodes to the cluster to accommodate increasing amounts of data.
- High concurrency. Handle more users at the same time, as your data is spread out across multiple machines.
- Extremely high availability. With more servers in the equation, ensuring uptime becomes that much easier.
Learn more about Active Everywhere databases in our ebook, The Power of an Active Everywhere Database.