Modern systems must run 24/7 even when hardware fails or maintenance is performed. This makes business continuity expensive and full of compromises like: buying additional machines that might not be as performant as your primary ones, setting up replication mechanisms to keep data synchronized and watching dust collect on an expensive disaster recovery setup.
Redundancy is also business critical when it’s time to run additional applications on top of your database – e.g., for reporting and analytics. However, redundancy is costly and requires compromises too.
NuoDB lets you build a global data management platform with guaranteed high availability, rolling upgrades, built-in data redundancy, and disaster recovery. With its hybrid cloud capabilities, it is able to store your data anywhere - locally, remotely, in a datacenter or on a public or private cloud. Store data in whatever storage system is most appropriate - on a directly attached file system, a local key/value store, or on a cloud-based storage service.
It provides both an administrative user interface and a scriptable interface. Both are able to dynamically provision additional resources when demand exceeds capacity or when hardware fails. The very nature of the NuoDB distributed architecture means no single point of failure and therefore zero downtime – made possible through a design we call the Durable Distributed Cache (DDC).
Key Features of NuoDB Continuous Availability:
- Full redundancy control
- No single point of failure
- Rolling upgrades
- Online schema evolution and backup
The DDC provides multiple points of redundancy: when the database observes a disconnect from a failed process, or a host that goes away, or a host that’s taken down for service, or when the network cable gets pulled out, the database simply reconnects to another transaction element, called a Transaction Engine (TE), within the NuoDB database. Simple, but, incredibly powerful because accidents do happen and maintenance tasks must be done.
Because the NuoDB architecture is so different from traditional DBMS architectures, continuous availability is a reality not an aspiration. It provides peer-to-peer, on-demand independence that yields high availability plus low-latency and a deployment model that is easy to manage.
Another benefit of the NuoDB architecture is the ability to make rolling software upgrades on an active database without taking it offline. This is a straightforward process in which a database administrator shuts down a NuoDB peer process, upgrades the one that has been shut down, brings it back up, and then goes on to the next process. When the last process running the old version of the software goes away, suddenly each NuoDB database process will kick to the next version, and begin using the upgraded version capabilities.
Rolling upgrades of server hardware are also supported. All you need to do is install and run new peer processes on your new hardware, wait until NuoDB has automatically synched its data, and then shut down the processes on your old server machine. This is all accomplished without bringing your database down.
Everything in the NuoDB system is also version controlled. If you want to add or drop a table column from a running database, NuoDB allows you to do that thereby giving you yet another powerful feature to keep your system up and running continuously.
Similarly, if you want to back up your running database, you can do this easily by adding a storage process, called a Storage Manager, to your running NuoDB database. Any time you are running a redundant Storage Manager with journaling enabled, you are getting automated replication.
NuoDB thought of all the angles on continuous availability from the architecture up, so you don’t have to.