How To Handle Sensitive IoT Data Across Your IIoT Infrastructure To Prevent Leaks Or Employee Misuse
January 9, 2018, by dpuron
As IoT devices continue to proliferate across industries, more and more diverse data is flowing from devices to decision support systems without human intervention or oversight. There are many business benefits when adopting IoT solutions in your industry, but as always, great power comes with great responsibility. One of the main responsibilities worth mentioning is the […]
As IoT devices continue to proliferate across industries, more and more diverse data is flowing from devices to decision support systems without human intervention or oversight. There are many business benefits when adopting IoT solutions in your industry, but as always, great power comes with great responsibility.
One of the main responsibilities worth mentioning is the proper handling of sensitive data. This is not only for your peace of mind but also due to increasingly stricter regulations. If you live or operate in the EU, I am sure you have heard about the upcoming GDPR, and the ePrivacy Regulation. These laws have to be thoroughly reviewed and planned when you plan your IoT deployments, be it commercial IoT or Industrial IoT (IIoT).
The purpose of this post is not to deal with the complexities and arguments on the impact of GDPR and other regulations on IoT. It is a matter of hot debate with quite divided opinions, and honestly, it looks like the main impact should be on the consumer rather than on the industrial space. The goal is to focus on how to properly plan, design and secure your sensitive data. After all, if you are doing an IIoT deployment it is because somehow you care about that data… and if you care, your competitors might care as well.
In order to illustrate our ideas, let us use an example scenario for an Industrial IoT deployment:
- Let us assume that we are a manufacturing equipment supplier with an established customer base.
- We believe that we can improve our customer’s operations through a proactive model consisting on monitoring our equipment operational data, analyzing it in our servers and extracting recommendations for our customers on how to configure and tune the equipment to improve the performance/output / KPI of that stage of their manufacturing process.
- Our Engineering staff has a clear idea of which data-points are needed to be monitored in each piece of equipment and our R&D team has developed proof of concept algorithms to identify recommendations for each case.
- Our Sales team is excited as this service would provide us with a competitive edge against our competitors and be the key to increase customer loyalty, let alone the recurring cash flow opportunity that comes with it.
This looks like a no-brainer to our executive team, and they are ready to provide the necessary resources for this new strategic project… until the Compliance Officer raises the question of how we would ensure we do not breach any confidentiality and sensitive information agreements we already have in place with our customers. Indeed, there are many ways this initiative could go wrong for the company. If we have any sensitive data leak or misuse, what looked like a sound competitive advantage, can also scare away our customers if they believe their data will not be safe with us.
For this case, we might think of a classic IoT architecture like:
In order to ensure that all the sensitive data concerns are properly addressed in your solution, you should at least consider the following activities in your design:
- Define what comprises data you care about, your sensitive data. Not all data are equally sensitive, but on the other hand, you would always be safe if you apply the maximum security to your deployment. Sometimes we tend to think that security comes with a price and complexity tag, but in reality, if we consider it at the design phase, you will find that you can secure your deployment just by making the right decisions from the beginning. Our recommendation for IoT deployments is that we set an overall level of sensitivity for your deployment, or at least per type of device you interconnect. This will help you decide which security measures to apply from the start.
- Apply a reasonable security level for the “level of sensitivity” identified and define what minimum security mechanisms must be in place. For example, for highly sensitive information (Machine and customer identification, key manufacturing parameters, etc), you need to secure the data at the origin. This will prevent both data thefts and misuse at the point of capture (i.e your IoT module in your machines).
For other ambient data, you might not be that concerned so applying a secure communication approach could suffice, (i.e: gateways in your deployment might not need to be secured at rest if the data they manage comes already encrypted from other devices, and they just take care of protocol transformation or connectivity.
Additionally, you should strongly consider keeping the firmware of all your connected devices properly updated, as you would not want to have any component of your architecture vulnerable to any known or unknown security threat.
[We also wrote down some tips to prevent your company being hacked that you might find interesting]
- Plan your deployment holistically. Typical IIoT architectures include interfaces with Cloud services for storage, analysis, and presentation of business insights out of sensor data. Ensure your choice of Cloud Services and APIs include security considerations (Secure communications enforced by the server, TLS/DTLS, Authenticated APIs, etc) and ensure you are using the “secure” option to connect to them (ie: some people develop using a less secure option and then forget to switch to the secure option for production operations). Often cloud APIs are the weakest link of IoT deployments, especially if they are custom made.
For your connectivity and cloud services we will always encourage you to use the maximum level of security you can afford, as changing the Cloud component of your deployment just might not be possible or affordable once the system is up and running, while you can add more data points (with potentially highly sensitive data) down the line.
One typical example of not planning holistically is when your IIoT deployment is perfectly secure but you hire a third party to develop a mobile app to interact with your devices. The third party development may not follow the same security practices. Hence, always perform the necessary pen-test and security audits to every component involved in any way or another with your IIoT deployment.
- Train your staff regarding the misuse of company and sensitive information. Remember that most of the data breaches come from human errors or misconducts, and most of them can be tracked to a lack of training or information about the impact of careless handling of the equipment.
In BarbaraIoT we are aware of the IoT hype, with all the promised benefits for all kinds of business, but we also know the challenges that impair its faster and widespread adoption. Most companies have smart ideas like the one described in this article, but when they face the challenges and potential pitfalls, they prefer to postpone or de-prioritize their implementation.
That is why we have conceived the Barbara Platform. We believe in our approach to IoT based on securing data at the point of collection through Barbara OS with configurable features like:
- at rest encryption
- secure communications
- over the air secure updates
- health and threat monitoring capabilities
- physical tampering proof mechanisms
will give you peace of mind that your sensitive information is well protected.
Article written by Pablo Portero, Co-Founder and Member of the Board at Barbara IoT.
Suscribe to our news letter to stay on top of the Industrial news and IoT trends