That one time you weren’t able to recover the set of emails that were a request due to legal hold proved painful. You now take backup exceptionally seriously. So much so, you now have hundreds of TB’s of data sitting on archive media. Economic advisors call it the new oil. So much insight could be made available via that wealth of data; only if there were an easy button to ingest what’s an organic, unstructured data lake. Enter the new era of backup companies such as Cohesity.
Nutanix-like Approach to NAS
Cohesity presented at Storage Field Day 18 (SFD). By now we all know the story of these scale-out HCI-like backup solutions. DellEMC’s Data Domain pioneered the scale-out backup appliance market. The likes of Rubrik Inc and Cohesity look to not just evolve the market but revolutionize the core capability. Cohesity’s CEO Mohit Aaron is a co-founder in HCI pioneer Nutanix. So, the ambitions are apparent.
Cohesity embraces the concept of being not just backup but secondary storage. The company is more than proud to display its scale-out NAS capability – the ability to add network storage by adding additional Cohesity nodes. The company is positioning the solution to take advantage of the massive amounts of data available in backup and archives.
Machine Learning Use Case
Think what if you could deploy some data combing agents on the nodes storing your backup data. These agents would identify patterns in the data as defined by a Machine Learning (ML) algorithm. The agents would push the data to an in-memory database such as SAP HANA. Once in HANA, data scientists could ask, in real-time, intriguing questions that inform better business decisions.
At Storage Field Day 18, Cohesity moved one step closer to that core capability. The company launched an App Store where application developers could deploy such solutions. The use case I laid out wasn’t too apparent given the presentation at SFD.
The delegates challenged Cohesity on the idea of running applications on secondary storage. Industry veteran Ray Lucchesi consistently challenged Cohesity on the viability and practicality of running applications on relatively pedestrian hardware.
I agree with Ray; there’s minimal appetite to run Tier 1 workloads on infrastructure designed primarily to perform backups. The resource contention alone represents a formable challenge. However, the use case above doesn’t require a tremendous amount of compute power. Also, Cohesity’s platform runs in cloud providers. I can imagine a use case for having these data lakes on a near-line solution such as Faction or ClearSky Data. Once the data exists close to the cloud, Cohesity could orchestrate much more powerful applications.
We are extremely early in this experiment. An obvious question, why not go cloud-native? Why not build the capability using AWS services? Cohesity’s primary advantage is the metadata from the backups and the ability to look into backups. I don’t believe anything precludes developers from writing applications that leverage cloud-native services using Cohesity API’s to scrap the unstructured data and metadata existing in backups.
I believe the Cohesity application store is an interesting take on the emerging market for these data protection companies. I’ll continue to watch the space closely.
Disclaimer: Gestalt IT, Cohesity, and Rubrik Inc. are past clients of The CTO Advisor content marketing services. No consideration received for this blog post.