Securing the Digital Restaurant at the Edge

The concept of smart kitchens is starting to creep into the restaurant industry.

Whether it’s adopting cook and hold ovens that reduce energy costs or adopting food safety management systems like Navitas, newer technology is being introduced to help restaurants run more efficiently and of course produce more revenue.

Food waste of course is one of the main problems being attacked by newer technology since it can cost commercial kitchens between five and 20 percent all food purchased.

While all these advances are great and can help increase revenue while lowering costs, they also create other problems.

Even something as simple as feeding in data from thermometers attached to fridges and freezers and providing a continuous feed allows restaurants to know when equipment breaks down before the food spoils. Likewise, being able to measure consumption automatically produces more accurate orders. Capturing POS order feeds in real time and correlating to how much is actually in the fridge allows forecasting models to be generated.

A lot of these sensors are small and simply shoot off a continuous data stream over Bluetooth or WiFi to one or more computers in the back of the house. From there more interesting software can collect and process the data and make decisions.

Even machine learning software, a newer style of programming that deals with large datasets are starting to find their way into the restaurant. Some of you might wonder why this is not just going straight to “the cloud?” There are a few factors at work here – one the concept of latency “how fast can you talk to the cloud” is a problem if you need a decision on something immediately – as in “When do I switch this oven off?” or “When does the temperature change to 92?”. This is usually coupled with the second problem of having very large bandwidth needs – e.g.: “how fat of a pipe are we shoving all of this data into?” When you have a multitude of sensors continuously feeding data back this bandwidth problem becomes larger.

edge

While all these advances are great and can help increase revenue while lowering costs, they also create other problems. Questions such as: how do you properly secure the computers and more importantly how do you actually manage them across your fleet of stores, take on greater significance.Even if you only have a handful of pizzerias in a small geographical area you are already dealing with more physical locations than your average Silicon Valley company has to deal with so securing and managing them become a task in and of itself.

McDonalds for instance has 14,000 locations in the US and 35,000 worldwide. DineEquity has about 1,900 Applebees and 1,650 IHOP locations. Now contrast that with the 56 total datacenters that Amazon Web Services, the dominant cloud provider, currently has. While they obviously eclipse the number of servers per location restaurants eclipse them on the number of locations themselves. Managing all of this becomes a very different task especially since there isn’t an army of 20,000 engineers working on it like the big FANG companies. The colloquial term for deploying servers in situations like this is being called the edge.

Chick-fil-A recently started introducing their edge server equipment to their chain of stores, so they could measure the overall efficiency of operations. The ability to know how many fries they should be frying at any given time or what the day’s demand for chicken might be can all be controlled and predicted when data is collected, and software is applied. What’s better is when it’s applied across the entirety of your store fleet you start getting insights you simply could not achieve before. While this is a newer practice you’ll start to see this everywhere in the coming years.

When you install smart kitchen hardware, remember the smart part always comes with software and that needs to live somewhere. Commonly, operating systems such as Windows and Linux are prevalent, but they are notorious for causing a lot of security problems. However, a third option known as unikernels also exists. Ask your vendors if they support that as a deployment option. Being self-contained they prohibit applications from being used as landing pads for hackers that might just be a hop, skip and a jump to things like kitchen display systems or worse your QuickBooks or timeclock system.

It’s hard enough to secure one computer in one location. When you have many locations, it becomes tougher. Just as you install security cameras, put locks on your doors and employ heavy duty safes you also need to secure the digital infrastructure you have, and the reality is that many systems are simply wide open.

Your restaurants are going to reap big rewards for taking advantage of new smart kitchen equipment but keep in mind that digital brings its own security challenges and the edge is uncharted territory.