Every customer’s storage requirements and deployment scenario are unique.  That’s why we designed our ActiveScale™ object storage system with a wide variety of deployment options to best fulfill customers’ diverse requirements of density, reliability, and data durability.  Today, I’m excited to introduce a brand-new option that gives our customers more flexibility, improved reliability and better performance than ever before.

Taking High Availability to the Next Level with Rack-Level Failure

The ActiveScale P100 and X100 systems are already available in multiple configurations, including single site, two-site with asynchronous replication, and three-site with data spread.  ActiveScale has always provided superior reliability, with built-in fault tolerance for the disks, servers, power, and network as standard features.  But now we’re taking that reliability to the next level with rack-level failure protection for single-site deployments, allowing ActiveScale to survive an entire rack failure without any disruption in service!

Re-Imagining of our Traditional “Geo-Spread” Deployment

This new architecture is a re-imagining of our traditional “geo-spread” deployment, which distributes data across three geographically dispersed data centers.  We have extended this model to include single-site deployments in order to provide rack-level fault tolerance within a single site.

Now, customers are able to locate the ActiveScale racks in separate power zones within a data center, close to data generators, or adjacent to other storage solutions.

With our synchronous data spread algorithm, your data is immediately consistent on all 3 racks once the write operation is complete.

How Many Failures Can it Sustain While Still Protecting Your Data?

Coupled with our storage policies, this allows an ActiveScale High Availability deployment to maintain data availability in the following failure cases:

  • 1 Full Rack Failure + 2 Disk Failures
  • Up to 6 Storage Server or JBOD Failures + 2 Disk Failures
  • Up to 8 Disk Failures in Up To 8 Enclosures
  • Up to 3 Metadata Server Failures
  • Front End or Back End Network Failures
  • Routing Device Failures

3x Performance Increase Over Traditional Single-Site Deployments

Another benefit to this approach is improved performance.  Unlike the geo-dispersed model, this type of deployment is not constrained by WAN bandwidth or latency and provides customers with nearly 3x performance increase over traditional single-site deployments.  Performance was measured at over 16GB/s on write and 19GB/s on read with 1GB objects across 500 concurrent threads!

This kind of throughput allows customers to apply S3 storage to a much wider range of use cases that also need massive scale, extreme data durability and high system availability that ActiveScale provides.

Working Hard to Protect your Data, So You Don’t Have To

Data has become more valuable than ever before, and we understand what that means for the challenges of managing it. Our engineering teams are working hard to ensure that your data is always protected and that our systems provide the highest availability possible.

If you’re looking for solution that can store massive data (scaling to over 60PB raw[1]), is extremely durable and available (with up to19 nines data durability), and delivers highly performant on-premises cloud object storage – ActiveScale is the solution for you.

Learn More


[1] One TB equals one trillion bytes and one PB equals 1,000TB when referring to storage capacity. Usable capacity will vary from the raw capacity due to object storage methodologies and other factors.

David is Sr. Technical Marketing Engineer for Western Digital, and has over 15 years of experience in software/systems/storage engineering.