Active-Active Database

Today, SaaS and enterprise applications, and their users are global. These applications require local performance, data consistency and resilience to data center failure. NuoDB resolves these challenges by running a single, logical, active-active (and even an active-active-active) database across multiple data centers with multiple master copies and true transactional consistency.

Since it can run simultaneously in multiple data centers (regardless of location or type - private, public, or hybrid cloud, for example), the NuoDB active-active database lets you build a highly responsive, geo-distributed database for continuous availability across regions with exceptionally low latency.

NuoDB’s architecture is based on a Durable Distributed Cache that consists of a network of peers with in-memory caches. These peers can be distributed across multiple data centers yet still operate as a single database. As a result, you bring the database closer to your customers, giving them faster response times. This effectively gives you an active-active database deployment that simplifies administration and eliminates the time your IT staff will spend on complex tasks like replication, backup and recovery schemes.

Key Features of NuoDB’s Active-Active Database

  • Transactions are fully available and consistent in all regions
  • Application developers and administrators view a single, logical database
  • Optimized for low-latency transactions over high-latency infrastructures
  • Commit protocols can be tuned by region
  • Built-in data redundancy across regions
  • Automatic, on-demand data replication
  • Detect and react to network partition failures via simple policies:
    • Full minority of selected region failure
    • Notify administrators that network partition has occurred

Unlike the traditional Master-Slave and Multi-Master approaches to distributing a database, NuoDB can scale out across multiple regions while remaining a high-performance, low-latency and truly ACID database.

For example, typical deployments have a decent amount of locality in terms of the data that is accessed. With NuoDB, data is only brought into memory when requested, and the full database is made durable by storage peers (Storage Managers in NuoDB). Transactional peers (Transaction Engines in NuoDB) naturally build up on-demand caches with affinity to the data and access patterns of a given location. Thus, communication is mostly localized, and there is little need for high-latency synchronous communication between data centers. If another Transaction Engine in another region needs to reference that same object, only then will it be replicated. This approach, and NuoDB’s use of multi-version concurrency control (MVCC), means you will always have a consistent view of the database.

Because NuoDB runs asynchronously in a peer-to-peer manner, it is able to avoid latency performance problems between remote locations. In contrast, the Multi-Master replication required for geo-distribution on traditional databases is costly, complex to configure, susceptible to data inconsistencies, and high-latency.

This all happens automatically “under the hood.” You don’t have to think about any explicit replication scheme, and you don’t ever have to think about active vs. passive, because NuoDB is always active, always consistent.

Recovery is also a natural feature of the NuoDB architecture. When NuoDB is distributed across two separate geographic regions in an active-active database configuration, then, should either region fail completely the other region will be able to take over its work. NuoDB will detect when a region has disappeared, and it will automatically connect the applications that were operating at the failed site to the operative site.

Modern applications and their users are distributed by nature. What they need is a single database running on a group of database servers in multiple data centers (or cloud regions) to give them higher performance, data center failover and the potential to manage issues of privacy and data sovereignty. NuoDB’s innovative distributed database architecture delivers geo-distribution without compromise.

Simplifying DR Across Data Centers

Simplifying Disaster Recovery across data centers

Our customer is a successful supplier of software services that are white labeled by network operators in North America, South America and Europe. Their main product has been very successful for its customers, based on dedicated equipment, installed in operators’ data centers. 

Read the Case Study

Want to learn more?

Check out these resources to learn more about NuoDB:

Visit the Product Overview Gartner Magic Quadrant Learn how NuoDB works