WekaIO
Disaggregating Performance from Physical Storage

While scale-out NAS systems revolutionized data centers in the mid-90s by quenching the ever-growing thirst for capacity, they failed to show grit when questions were raised about better performance in the post-millennium period.

Today, administrators are demanding storage requirements based on their application needs, and not the other way around. The traditional shrink-wrap storage systems that force fit applications to run on a rigid platform are no longer relevant for the cloud era.

“In the first generation of hyper-converged environments, it was about sharing the underlying storage infrastructure and not about sharing the data,” says Barbara Murphy, VP of Marketing at WekaIO. “At WekaIO, we are pushing the virtualization processes one layer further. We’re not just providing the ability to virtualize physical storage, but enabling application data sharing so that performance and data accessibility are never compromised.”

Murphy explains that in the Web 2.0 era, users expect their cloud environments to deliver the same performance and capabilities as their on-premises infrastructure, and data should move seamlessly from an on-prem to public cloud environment. Data storage services and application performance go hand-in-hand, and cloud users will want the ability to provision the quality of service applications demand.

WekaIO’s platform is ideal for environments with shared storage needs and multiple users that require access to the same data sets, as opposed to simply serving virtualized environments with fixed bandwidth-centric or IOPS-centric configurations. WekaIO meets customer performance service levels by dynamically assigning sufficient compute resources as needed for the duration of the application workload. Once the application workload is complete, WekaIO returns the compute resources back to the shared server pool for other applications to use.

“Storage performance and capacity are completely independent in the WekaIO model. By understanding the amount of compute resources available to WekaIO for performance, we can predict in advance the application performance our customers are going to see,” explains Murphy.

Such preemptive capabilities go a long way in addressing certain niche enterprise requirements.

WekaIO provides the ability to virtualize physical underlying storage as well as enable application data sharing, never compromising on performance and accessibility

Legacy infrastructures cannot discriminate between two users— one with millions of small files and negligible bandwidth requirement, and the other with large files requiring high bandwidth and low IOPS.

One customer from the media and entertainment industry consolidated data across 80 individual storage appliances into ten high performance servers by leveraging WekaIO—illustrating a use case where a customer could realize the true potential of cloud technology. The defining factor of their success story was the architecture— something one would call “fast-and-deep.” Applications with low latency and high performance IOPS run hyperconverged on the “fast” application tier, while the S3 cloud backend delivers “deep” long-term storage at the scale that you would expect from major cloud infrastructure delivering high bandwidth.

The shrinkage of the physical storage footprint dished out several benefits— reductions in power consumption, cost of operations, cooling requirements and most importantly, physical floor space, which was a big financial burden on the Southern California based company.

According to Murphy, the cloud industry is heading toward a scenario where the public-private cloud delineation will wane. There will only be a compute environment, where potentially some of the data will exist on a public cloud, and the rest will reside in an internal one.

WekaIO provisions application services, as opposed to storage services, and helps organizations achieve their performance needs without the limitations of their current storage environment. “Our product is both cloud-native and runs in-house,” summarizes Murphy. “WekaIO combines those two in a way that it doesn’t matter to the user where the data and applications are running.”