Data storage considerations for a DevOps environment
A comprehensive collection of articles, videos and more, hand-picked by our editors
Supply chain standard organization GS1 Canada moved into hyper-convergence on the data protection side -- beginning with its DevOps storage -- before taking the plunge with primary data.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
GS1 is a nonprofit group responsible for bar code standards, with about 20,000 subscribers. The organization assigns product identifiers (GTIN, UPC) for bar codes and manages a product registry, requiring it to move large volumes of data between manufacturers and retailers.
Sase Janki, GS1 Canada's senior vice president of technology, said the Toronto-based organization has about 200 TB of data, and it is growing around 30% to 40% per year.
"We are in the infancy of a data explosion," he said.
To get a handle on that growth, GS1 became an early customer of startup Cohesity Inc., which describes its appliances and software as convergence for secondary storage, including backup and archiving.
GS1's storage infrastructure is divided into DevOps and production data. DevOps handles testing, production support and quality control. Production data includes an external-facing, web-based application that manages synchronization of data bar code information.
The organization uses NetApp for primary storage. It replaced its backup software -- mostly Symantec/Veritas Backup Exec -- and direct-attached backup hardware for DevOps with Cohesity.
"We weren't able to keep up with the volume growth," Janki said. "All of our suppliers care about e-commerce and wanted to get product data on our website. There was a huge push to get reliable data up. We had challenges in recovery. We could back up, but never knew when we could recover. We couldn't predict our cost of storage.
"I wanted a simplified solution, not a complex solution where you had to configure the network, configure the infrastructure and configure backup jobs."
A services provider pointed Janki toward Cohesity when the vendor was still in stealth. Janki was sold on the concept -- a 2U appliance with hard disk drives and solid-state drives running Cohesity Data Protect and Data Platform software. Cohesity's goal is to converge all secondary data on its platform.
"I said converged storage might be attractive to us," Janki said. "It's all centralized for us. I don't have to worry about all the configuration and work that goes into it, because I don't have the staff for that.
"Two things about Cohesity caught my attention: Hyper-convergence of everything in a box was appealing, [and] so was the speed of backup and recovery. I had to change the way we did storage and figure out how this was going to be easy for me."
GS1 brought Cohesity in for a yearlong trial beginning in December 2014. Janki said the performance boost with DevOps data led to his decision to begin moving his production virtual machine data to Cohesity in 2017.
He said the DevOps team has about 60 TB on its Cohesity box, and he added a second unit at a remote location for failover. Cohesity software allows developers to clone the latest backup of their production application stack and run it directly off the Cohesity Data Platform to eliminate redundant copies.
Janki said he thought using Cohesity with DevOps first would provide a good test to see if it can work with a large, unstructured data store. "DevOps is important, but if something goes wrong in DevOps, it doesn't impact my partners," he said.
He said things have gone right to the extent that he is ready to use Cohesity for production data.
Sase Jankisenior vice president of technology, GS1 Canada
"Recently, we had an issue where we were forced to use Cohesity for production servers and had to recover, and it worked," Janki said. "My confidence is there. My objective is to scale it up to a full-scale production environment. I wouldn't have invested and gotten to this point without intending to get to the next level."
He said his Cohesity box also improved speed of backups considerably, and it is easier to set up and use than his production backup and storage.
"We have a reporting and diagnostic UI that gives us information we didn't have from Backup Exec," Janki said. "Everything is in a single box, so we don't have to set up network connectivity the way we have to with NetApp. With NetApp, we have to bring in a specialist. We set up Cohesity ourselves."
Janki said he is interested in taking advantage of Cohesity's public cloud support, which was added in April.
"I'm looking to see if I can reduce costs for power and space and use a hybrid cloud setup," he said. "We're doing that now in DevOps. We want to get more confidence [before using the cloud for production data]. Then, I will see if I could use it in a more direct manner for our actual production system."
Now that he's sold on hyper-convergence, Janki said he is considering it for primary storage.
"I think I can benefit from a hyper-converged infrastructure for primary storage," he said. "But I have to rebuild my production environment to do that."
Value in merging backup and archiving
Hyper-convergence products vary for data protection
Explore how backup is converging with other techs
Who won at the Best of VMworld 2016 awards?