Scale-Out Performance

Traditional Database Architecture

Traditionally, IT organizations support business-critical applications using "scale-up" approach:

  • Buy the biggest server you can afford, sized for PEAK capacity "just in case"
  • Add another server the SAME size in a High Availability cluster for fault tolerance.
  • Add another server the SAME size at a different site for Disaster Recovery.
  • Add more similar servers for Development, Testing, and Support.

The problems with the Traditional approach:

Wasted Resources: Big SMP servers are often "super-sized". The HA Cluster's secondary server is typically idle. The DR site is also sitting idle.

High Cost: This approach requires expensive servers, multiple HBA interface cards for each server, long distance third-party replication services and a big WAN pipe.

Points Of Failure: After all the money you've sunk in, you still don't have continuous 24x7 availability.

xkoto gives you a better approach. Light up all your assets - from your HA cluster to your DR site. GRIDSCALE's patent-pending active-active technology will give you the scalable performance and continuous availability you need.


Achieve Scalability with Database Virtualization

With GRIDSCALE, you can virtualize your database environment. Typically our customers start with three or more commodity servers or virtual machines (VM) and use GRIDSCALE to distribute the workload across servers in the cluster.

Add servers to the cluster for scalable performance - the servers are all active all the time (no master/ standby). Later, add another server at a different site for disaster recovery - even the disaster recovery (DR) server is active.

The results:

High resource utilization: All servers are active so the read load can be spread out, providing horizontal scalability.

Low cost: You can use commodity servers, local disk storage, and standard ethernet.

Resiliency: Now you can deliver continuous (not just high) availability, rolling upgrades, high fault tolerance, automated resynchronization, and distribute your data safely across multiple sites.