What happens in the Edge, stays in the Edge - Qwinix.io

 

Qwinix
Blog

What happens in the Edge, stays in the Edge

August 19th, 2019 | Tech

A Qwinix QuickTake with CTO Leo Murillo

Edge Computing and Edge Security

As in Vegas, what happens in the Edge, stays in the Edge

Just as you may want details about your Vegas vacation kept within city limits, edge computing keeps sensitive information private within the boundaries of the edge.

 

How Does Edge Computing Keep Data Safe?

 

Let’s look at an example. You have some data stored in-house which consists of pet images. You want to process those images and identify how many are cats to determine when you need to add more cat food inventory to your local stores. You plan to leverage Google Cloud Platform to manage secure machine learning architecture. The catch is that your company has a data security policy that says the images can’t leave your company devices. What do you do?

 

When you want to keep data private, utilize the edge. Edge devices can be your own data center’s edge servers, devices of users in your office, or off-premise IoT devices using Raspberry Pi in trucks, warehouses, homes or many other places. Using Google Cloud Platform and edge devices, you are not limited to any particular location for machine learning. You can choose your machine learning location for reasons such as architectural efficiency, data consistency, and data privacy. Google Cloud Platform gives you a common tool chain you can use to run machine learning anywhere, using edge computing.

 

In this example, you feed your pet images through an initial iteration of a machine learning algorithm on IoT edge devices, which can be your company’s own secured IoT edge devices. This output is then submitted for machine learning processing on your secured edge servers in your data center. At this point, your data has not left your company’s devices and is still personally identifiable.  By keeping some machine learning on-premise on your edge servers and devices, you keep your sensitive data private and contained within your domain. 

 

After the data is processed on your edge servers in your data center, the abstracted, obscured data is shipped to the Google Cloud Platform for more in-depth machine learning processing.  In the Google Cloud, you can do your bulk machine learning processing using your obscured data.

 

Taking it a step further, you can use the cloud edge as the outer layer of your neural network. Then the layers further inwards are handled in the Google Cloud. What those layers receive is not raw data, but already processed, abstracted data.

 

Edge Computing and Google Cloud Platform: Run Machine Learning Anywhere

 

Using the edge and Google Cloud Platform, you can choose where you do machine learning, whether it be on edge IoT devices located anywhere, on your edge servers, or in the public cloud. The Chrome browser can build and train machine learning models using a JavaScript library built for TensorFlow. The benefits of edge computing are enhanced with Google Cloud Platform, allowing you to customize where machine learning takes place based on your data privacy and processing efficiency needs. Using Google Cloud Platform, you do not have technological constraints when choosing a location for machine learning. Google Cloud Platform gives you a toolset for machine learning and edge computing you can run anywhere.

For more information about how edge computing can benefit your company, please reach out to us.

Submit a Comment

Your email address will not be published. Required fields are marked *