Dutch Food producer, Aviko, creates a high-availability cloud setup to support 24-hour operations

“Implementing a new SAP system was never a problem. Our challenge was to guarantee the system would never go down.”

Rene Aalderink, Cloud and Connectivity Engineer at Cosun Shared Service Centre (Cosun SSC) is talking about the importance of a high-availability setup and how it has helped his team implement a new automated SAP system for the Dutch food processing company, Aviko.

Part of the Cosun Group, Aviko is one of Europe’s largest producers of fresh, frozen, and dried potato products. At the center of its supply chain is the company’s cold store warehouses where products are kept between production and delivery—a core aspect of the business that Aviko wanted to automate by implementing a new SAP system to remotely control each warehouse in a standardized way.

It was a digital transformation that fell into the hands of Cosun SSC, the team which manages the IT infrastructure for every business within the Cosun Group. But with the fast-moving nature of these cold store warehouses, Aviko couldn’t afford any break in its production line, giving Cosun SSC just four months to implement and test this new integrated system.

“Our team has been running SAP systems across the Cosun Group for over 10 years. But this new system needed to run 24/7 without any downtime to keep the warehouses running efficiently, which gave us a higher availability number than our current data center could handle,” says Aalderink.

“The new system also only ran on Linux, meaning we needed to create a new IT landscape to support the system and have it all implemented in record time to meet our December deadline.”

Innovating the frozen food world

Since its incorporation in 1962, Aviko has grown to become one of the four largest potato processors in the world. Its supply chain operates 365 days a year, from farming 1.8 million tons of potatoes annually to shipping products to more than 110 countries.

Before reaching the plates of its customers, every Aviko product is sent to one of the company’s cold stores, a seamless operation where pallets are logically arranged across the warehouse to create efficiencies between production and delivery. The slightest delay or missed shipment could create a buildup of pallets, causing production to shut down and cost Aviko valuable time and money.

“Aviko has used SAP systems in its cold stores for a long time to communicate with third-party tools that manage the warehouse logistics,” explains Aalderink. “Until recently, this process was not very innovative. But in August 2019 that changed with Aviko’s new automated warehouse project—a plan to switch out the third-party vendors and run everything through a new SAP EWM system based on S/4HANA.”

Cosun SSC’s first option was to deploy this new workload as a clustering system, with multiple computers working together across an on-premises data center to support S/4HANA. This required upgrading specific hardware that the team didn’t have time to source or implement, leading them to explore running this new workload in a different way than other Cosun businesses—by deploying Aviko’s infrastructure on Microsoft Azure.

Launching the first fully automated warehouse

“We already had a base infrastructure set-up in Azure as an additional data center, so we did the numbers to see if it could support the higher availability SAP S/4HANA required,” recalls Aalderink.

“It was looking good. Migrating to Azure also offered other benefits over our on-premises data center, such as the possibility to start deploying right away by following Azure whitepapers and simplifying our implementation best practices—all factors that would help reduce delivery time and keep to our tight deadline.”

With the decision made, Cosun SSC began designing this new IT landscape with help from member of the Microsoft Partner Network, Sabaas BV, a team with 20 years of experience running SAP workloads and expertise in both Linux and Azure, which would be needed to implement S/4HANA.

“We started working in a sandbox environment to help test and fine tune the clustering infrastructure needed to run this high-availability setup,” says Sander van Gemert, Managing Partner at Sabaas BV.

“This was implemented in both Azure and Linux to give Aviko reassurance that if one server fails, the system will still operate like normal. At times it was a complicated process, but having the Microsoft team on board meant we could get engineers involved to help fill in our knowledge gaps around running these cluster set-ups in Azure.

“Their real-time assistance really helped speed up the whole process, and by November we were ready to move into production, ahead of schedule and still within budget,” van Gemert says.

Implementing the system is just the start

With the new S/4HANA system in place, Aviko was able to successfully launch its first fully automated warehouse in 2020. A big step from its on-premises SAP system—and an improvement, with data analytics to now plan as well as execute the entire operation.

“By combining the Azure connected SAP virtual machines with the S/4HANA in-memory database, we can now quickly access any necessary data to run our processes in a faster and more efficient way,” explains Kees Jansen, IT Manager at Aviko.

“It means operations like the Customer Order Decoupling Point can be completed in less time and placed later in the process. And by using Azure, we can also eliminate expensive capital expenditure costs and benefit, not only from the operational expense model, but also [improve] the flexibility it provides.”

Aviko’s new streamlined process begins the moment products are unloaded from the factory. Each pallet is scanned and transported by robots to over 40,000 storage places within the cold store, all pre-determined by the SAP system and based on when each product will be needed. When a customer order is placed, the correct pallets are collected and loaded onto trucks in the most optimum way, keeping the warehouse rotating efficiently without any downtime.

“It’s a fully automated system that calculates where pallets need to go and in what order, and then communicates those orders to each robot,” says Aalderink. “Even if a customer cancels an order last minute, the software can make the necessary adjustments to keep the warehouse in check, so there's a lot of logic behind it.

“This warehouse project is a new milestone for Aviko, and a major project for Cosun, to get up and running so quickly,” he adds. “The flexibility of Azure also means we now have a future-proofed system that won’t need new hardware again in five years.”

Aviko’s automation goals extend beyond this initial project. “We can just keep adding to the infrastructure and creating new workloads like our ARV project—a new autonomous truck that can load products at the factory and unload them into the warehouse. And we’ll be the first company to launch this type of vehicle in the Netherlands,” Aalderink continues.

Looking to the future of potato production

With this new, advanced cold store now in production, Aviko is looking to open two smaller warehouses in the Netherlands and a bigger one in Belgium that will all be controlled by the same central SAP system. This would be a digital step forward, and would help the company raise efficiency between production and delivery, as well as reduce operational costs by running multiple warehouses on the same software layer.

“This digital transformation has helped lay the foundation to unlock big data and use it to improve our integral KPIs,” says Jansen. “Whether that’s developing heat maps to provide insights into the physical movements in our warehouses or applying new techniques, such as IoT and machine learning, to further enhance our operation.”

“Right now, we are mainly focusing on the intelligence for moving and storing the pallets. But running S/4HANA on Azure opens the door to run all kinds of innovations in the future,” adds Aalderink.

“One day we could be monitoring the potatoes growing in a field and at the same time the condition of certain mechanical parts in the factory so that we know when they need to be replaced. All are big data [projects] that will help us create a more efficient ecosystem within every stage of Aviko’s operation.”