Adopting Hyperconvergence and Adapting to the New Realities in Retail
HPE SimpliVity
The retail environment is changing at an exponential rate. Consumers are always adapting to the world around them, so their habits are constantly evolving. As a result, retailers must constantly change course to meet consumer needs and continue to provide the exceptional service consumers have come to expect. Despite many retailers earning high revenues, profit margins remain razor thin.
On the surface, the retail sector may seem simple, but it has a highly complex infrastructure compared to other industries. We have a very distributed structure and have to manage these distributed and complex systems using fewer people and smaller budgets than other sectors. Finance, for example, can have budgets and IT staff four or five times the size of ours.
The Constant Challenges of a Retail Success Story
As an electronic engineer by trade, I’ve spent the past 20 years working in the retail sector. For most of that time, I have been working at Turkish retail company Migros Ticaret A.S., most recently in the role of systems and cloud infrastructure manager.
Migros has 2,200+ stores and 28 warehouses in Turkey, as well as stores in Kazakhstan and Macedonia. We have 12 million active loyalty cards. We use 2,400 suppliers in order to keep our shelves stocked.
To keep up with this growth, Migros has more than 45,000 employees to support all operations, from frontline sales associates to leadership to warehouse and IT employees. As an IT department, we are also an R&D center. Migros grew at a rate of 23.9% in 2019, with last year’s revenue totalling ₺23.2 billion (Turkish lira). And to maintain this rate of growth going forward, we had to overcome several key challenges within our internal systems.
Retail runs 24 hours a day, seven days a week. Retail storefronts are open to customers during the day while warehouses and logistics continue to operate overnight. As such, one important obstacle to overcome was the required planning of maintenance. Everything needed to be planned at least two or three weeks in advance.
Another critical issue was the need for at least four IT experts to manage our virtualization platform. Administration and operations were very complex, so when systems went down, needed upgrades, or required troubleshooting, IT found it difficult to find the root cause of problems. Finding and fixing issues required a lot more time and energy than was necessary, and the process was rarely smooth.
Five years ago, we decided that we needed to smooth out the process. We began the search for a simpler, more flexible solution that would allow us to run our operations while reducing costs.
The Hyperconvergence Takeover
There wasn’t much vendor diversity at the time we started our search for hyperconvergence systems. We had seen SimpliVity (now HPE SimpliVity) at international trade shows, but when we returned to Turkey to compare our options, there wasn’t a lot of choice on the market.
HPE SimpliVity and one more product were the two main hyperconvergence vendors. HPE SimpliVity had more features than the other, including backups, disaster recovery (DR), and fault tolerance. HPE SimpliVity was also a more cost-effective option—with the other, the replication and automation software we would have had to purchase for DR replication and automation alone would have cost us the same amount as the total project cost of HPE SimpliVity.
It was a clear choice. We replaced our traditional storage-server environment with HPE SimpliVity.
HPE SimpliVity continues to give us the ability to scale up in both storage and server virtualization. By getting rid of the management complexity of the traditional environment and replacing it with HPE SimpliVity, we’ve reduced operational risks, and boosted data reduction and simplicity.
Transforming Our Core Systems
We use HPE SimpliVity on Migros’ core systems: server virtualization, application virtualization, and desktop virtualization. In total, we have 45 SimpliVity nodes, 50 compute nodes, and 2,500 virtual machines (VMs).
All of Migros’ server virtualization environments run on HPE SimpliVity. All of our infrastructure runs on 1,500 virtual servers, including an active directory, SAP, Microsoft Exchange server, IIS, SQL, CRM, ERP, and more. We run HPE SimpliVity backups for all VMs each night, as well as remote backups for all of our VMs.
With HPE SimpliVity, we can now handle more than 12,000+ concurrent sessions on our application virtualization environment. Almost 100 virtual apps run on this environment, and nearly 500 VMs regenerate main templates on the system each night without any input/output operations per second (IOPS) issues. Our Application Virtualization System manages all of our applications, such as our store and office apps, browsers, ERP, weight machine apps, critical warehouse apps, and more. These virtualizations run on our PCs, handheld terminals, thin clients, mobile phones, and tablets, so it’s universally accessible for our teams’ needs.
In all of our warehouses, all of our thin clients make use of desktop virtualization. Up until recently, there were nearly 300 VMs running concurrently, each of which regenerated main templates on the system each night. Again, we don’t have to worry about any IOPS issues.
We’ve also begun to use HPE SimpliVity when refactoring our in-house applications with microservice architecture on our container platform, choosing DevOps tools that will run on HPE SimpliVity. It’s a big project, and the main ERP software will go live later this year.
Stress-Free Three-Second Backups
The server virtualization aspect of the system overhaul has been critical when it comes to disaster recovery infrastructure and procedures. Five years ago, we didn’t have any DR centers, and we needed one. Once that was created, however, we needed to replicate our virtual machines and write a DR scenario at the other site. We used RapidDR to replicate HPE SimpliVity to our DR center and write the DR scenarios over there.
Before HPE SimpliVity, we made the backups locally with a popular backup solution. This was ineffective and essentially inoperable for our needs because this backup took days to complete.
Sometimes this software couldn’t update the snapshots, which was a very big problem within our environment. We also tested a one terabyte VM backup, which took six hours to complete while requiring our team to monitor the process throughout. If something failed at any point, we would have to start the process over from the beginning.
With HPE SimpliVity, individual backups take three or four seconds, and all of Migros’ VMs—both local and remote—can be backed up in less than an hour. Backup operations occur during the day, and the efficiency of HPE SimpliVity’s backup process takes the strain off of our servers. It’s simple and efficient. Everything is automated, so we now have much more time for projects that add more value to our business beyond IT systems operations. This means that we can focus more on scaling our business, adding to our existing 2,500 VMs as needed.
We restore a VM once a month at most. Doing so takes a maximum of 10 seconds, regardless of whether the size of the VM is 100 gigabytes or 3 terabytes.
Doing So Much More with So Much Less
Not only are we saving time maintaining our new system with HPE SimpliVity, but we’ve become more efficient. Migros has improved our server virtualization efficiency at a ratio of 15.3:1, our application virtualization at 16.3:1, our desktop virtualization at 25:1, and our continual platform at 18.9:1, compared to our former traditional systems.
We’re also saving resources. Our traditional storage-server environment required separate teams to manage each of our six or seven separate systems. Today, just one IT expert is enough to manage and operate our comprehensive HPE SimpliVity system.
The system is suitable and scalable for big enterprise companies like ours. Since HPE SimpliVity runs on vCenter—which is an industry standard—it means our IT team doesn’t need to learn new steps to manage it all. It also means that, should we need to hire someone, they’ll likely already have the skillset needed to run it all. The rest of our IT staff can be redeployed to analyze and optimize our other systems, which is highly valuable due to the need to stretch our minimal resources as far as possible. They are using their skills more efficiently now.
Over the past five years, we’ve reduced our IT operating costs by about 50%. Some of that is because HPE SimpliVity is also a green IT system. Switching to HPE SimpliVity has resulted in a 70% reduction in energy usage.
Flexibility is a particularly important aspect of our retail business, especially given the current global realities that require the ongoing expansion of remote-working capabilities amongst our employees. This is a situation where desktop virtualization has been decidedly valuable. We were initially running 300 virtual desktops and needed to spin up 700 more to give our employees the flexibility to work from home during the COVID-19 pandemic. Using HPE SimpliVity, we were able to do so in just 12 minutes. HPE SimpliVity can handle these additions, while still managing the 12,000 concurrent stations running Virtual Apps in our warehouses, stores, and offices.
Managing the Present While Building for the Future
The retail sector changes very quickly and in our company, the IT department must be a drivable business. There isn’t much room for error if we want to survive, and much less room for error if we want to thrive. HPE SimpliVity is a horizontal growth architect, so we don’t need to upgrade any storage servers and the like. If we need to grow, it’s as simple as adding in a node.
We are growing fast, and hyperconvergence makes our systems less expensive and simpler to operate as we scale. HPE SimpliVity allows us to keep things simple, freeing up the creativity and budget we need to manage the present and plan for the future.
No one can say for sure what to expect in the years to come, but we’re confident that our systems and infrastructure will be rock solid, allowing us to adapt to the new realities of retail.