We sat down with Steve Knipple, Chief Technology Officer and Vice President of Engineering at EasyStreet, an infrastructure-as-a-service and managed cloud provider specializing in tailored IT-as-a-service solutions. As a specialist in corporate IT transformational programs, Knipple has over 20 years’ experience in IT enterprise strategy, architecture, management, and operations. He discussed meeting customer challenges and considerations for flash in his business today.
What is the mission of EasyStreet?
Our mission is to provide reliable, responsive, and responsible IT as a service. We build long-term relationships; prioritize quality of service over cost; and synthesize client needs with standardized architecture and processes for reliable, versatile services. We also give back to the community and evangelize sustainable practices to “do well and do good.”
What matters most to your customers? For example, improving service delivery, being more competitive, what comes to mind?
What comes to mind first is always high availability for mission-critical applications. Our clients want to know that their systems are available 24/7/365 – all the time, or what we call “always on.”
Also important is that they can talk to somebody. They want to be able to call and speak to a knowledgeable engineering person or a knowledgeable operations group to help them.
We do a lot of service integration and management consulting with clients. Many clients don’t know how to move to the cloud and want help to ensure a successful migration. Therefore, they see us as a trusted advisor to engineer, deploy, and maintain their solutions. In most cases, we are dealing with many steps to migrate and have to create a hybrid environment with the proper roadmap to meet their needs.
Recently you made a decision to put flash into your IT environment; what were some challenges you were trying to solve?
We currently run a mixed environment of storage, including SAS, SATA, hybrid, and all-flash. While hybrid arrays were working well in some environments, we found other shared environments where the mix of traditional disk was an issue for more performance-sensitive clients. Shared environments presented storage performance impact on SaaS and e-commerce clients where it was imperative that sites load faster and that complex transactional queries execute faster against large datasets. Enterprise clients also had performance needs for VDI workloads and batch reporting performance for systems such as ERP.
Therefore, as our client base sophistication grew, we started to see a trend toward an increase in IOPS-per-gigabyte requirements. Clients simply don’t have enough gigabyte storage requirements to make it cost-effective to put in all those extra spindles because of the IOPS. More clients needed three or four terabytes, and some, five or seven terabytes, as well as extremely high IOPS – mainly for database applications. We started looking at flash in the marketplace as the perfect solution to meet client demands.
Since you were already a NetApp customer, what led you to deploy the NetApp® all-flash array?
It’s refreshing to see established storage companies like NetApp that have all the processes and support mechanisms in place as well as really good customer service. However, while other companies are learning how to play in this new flash market, NetApp brought its flash portfolio and the EF540 to us, and it performed well. NetApp had the right people show up, who knew what they were doing. That was a big differentiator. That doesn’t always happen with every vendor.
We looked at around five to six flash array vendors; there are a lot of companies entering the space. We had four criteria components, which NetApp met:
Are there any customer flash use cases with the EF540 that you can discuss with us?
Absolutely. Let me discuss two, one on the database side and the other VDI.
The first client is a global retailer with its worldwide ERP hosted at EasyStreet. This client has a large, complicated supply chain with large batch jobs that run on regular intervals. They run weekly and monthly batch jobs for their supply chain, some taking as long as 30 hours to complete. (What I like to call a “global weekend.”) The problem required more horsepower to were optimize these jobs and to ensure their disaster recovery system was back up and running on Monday morning at 8 a.m. without any problems. We certainly helped them achieve those goals with NetApp. Their backup times were incredibly faster, with a reduction of 90%, and batch jobs execution times were reduced 20% to 50%.
The second client, a food manufacturer, has a large VDI deployment across their office space and shop floor. With a diverse and unpredictable server workload affecting their VDI load, this particular client’s older array was giving them performance problems. We moved their VDI load to a NetApp EF540. It was a pretty dramatic improvement for them. Login times were reduced by 80%, significantly improving the user experience.
With two EF540 arrays deployed, do you expect to have any additional all-flash arrays in the near future?
Absolutely, the capability of the EF540 helps us bring in new business, including some considerably large clients. All-flash in our service catalog has allowed us to have discussions with clients that we never could have had without it. Flash has made us more competitive, but, more important, our clients are happier as they now solve problems faster.
The Flash Memory Summit is happening August 5 to 7 in Santa Clara, California! Visit NetApp booth number 217 to see the EF540 and to meet with NetApp.
On Tuesday, August 5, from 11:30 a.m. to noon, Steve Knipple will join NetApp VP of Flash Products Ty McConney in his keynote, “Making Flash Work for the Enterprise Customer.” Also follow NetApp on Twitter, including conversations at the show and highlights of NetApp executive presentations.