DCSL (Private Cloud) – Iopstor streamlines IO latency issues for 500 users

Background
DCSL is a laboratory for grading and certification of diamonds, coloured stones and jewellery to more than a million customers. Due to the large operations their enterprise storage could not keep up with their virtual infrastructure. Users would have machines and software hanging and slowing down the processes there by affecting productivity. Their in-house ERP and database required IOPS with low latency which could not be supported by the enterprise storage in use. Their SQL database size was a couple of 100 GB’s and the SQL consultant mentioned that they require higher IOPS for fluent work operations. There was a lot of downtime affecting their productivity as their enterprise storage kept failing on a weekly basis actually stalling the entire business operations. They were advised to increase their server compute and memory and buy a new enterprise storage to resolve their latency issues. The cost of such hardware was exorbitant and unwarranted according to their IT Team. They started looking for alternate solutions and contacted us.

Solution
IOPstor understood their work flow and studied their entire infrastructure for a fortnight and came to a conclusion that it was a IO latency which was causing all the problems in the IT infrastructure. We recommended a change from Spindle hard drives to Solid state drive’s on their existing storage. Even with the upgrade of their enterprise storage to solid state drives the problem persisted. We finally convinced them for a POC of our IOPstor ZFS solution to resolve their roadblocks. They agreed to deploy our storage solution in their enterprise infrastructure for a period of 2 weeks on a trail basis. A 8 bay IOPstor storage solution was provided using the same SSD;’s from the enterprise storage with a RAID Z2 configuration.

Results
They were shocked to see the results within the first 2 days where there was no downtime or latency issue on the IT infrastructure. Our IOPstor storage solution was benchmarked to be at least 70 % more efficient and faster than the previous enterprise solution. Implementation of snapshots were created for a role back incase of any data corruption due to virus or ransomware. A replication storage at a remote location was deployed with continuous sync to the main HA , as a disaster recovery solution. Users were very comfortable with the speeds as the IO latency and network latency issues were resolved. There were able to process 15000 tally entries a day. The SQL database increased and yet they did not face any IOPS issues. They finally deployed similar solution in 4 other locations in India and since then have never faced any downtime or latency issues whatsoever.