On the first night of VMworld EU 2018, I attended a blogger reception for Cohesity organised by Patrick Rednap. I was trying to cadge a VIP ticket to the vRockstar party that evening. While there I met the who’s who in the blogger community, most people I knew for a long time.
But presenting was Rawlinson Rivera CTO for Cohesity. For me, I didn’t know much about them except they tagged themselves as secondary storage and usually, I avoid storage discussions and find them boring an old news.
Apart from Rawlinson’s big bravado and humour I wasn’t sure if I’d be even interested but It was worth the while. I found out a few interesting things about them. So this is what I picked up:
They abstract the data layer in their platform giving a consistent feel across private and public. So you get the same functionality in their services regardless of where you are sitting. Services like Data Protection, file&object and analytics. And its all API first… The problem they are trying to fix with this is Mass Data Fragmentation.
They’ve done some testing to prove they are Web-Scale up to 256 Nodes Tested ~5PB raw capacity showing Linear performance scaling using Sequential and random IO.
I asked the question are they doing dedupe client side and they claim to. They are doing Dedupe at every level – inline to Post job.
They acquired some tech called Helios and my understanding this is a manager of managers. Single UI for everything you need and offers some analytical functions.
One cool feature I found is they by default compress any video files, this is supposed to reduce backend storage needs like you wouldn’t believe. All policy driven.
One feature I’d have to research was a feature called Cloud Spin in which they are supposed to be able to help refactor applications by pushing data into a test/dev area without dragging along with it the whole OS etc…
Anyway, we had a bit of fun. Rawlinson a fun presenter and he caught our attention.