Change the default passwords! Know what devices you have and where they are! Get rid of old devices! If it can't be patched, yank it out! Those are some of my recommended best practices when it comes to enterprise IoT devices -- whether they're located in corporate offices, out in the field (from farms to mall kiosks), at customer sites or even hiding inside consumers (think medical devices).
Smart photocopiers, factory-floor systems and even handheld scanners used by parcel delivery staff seem like they're new, because of the name "Internet of Things." However, connected and embedded devices have been used for business purposes for decades, and they're getting new attention.
The advent of cellular data modems and ubiquitous. WiFi and Bluetooth have made it easier to deploy IoT devices. Cheap microprocessors, memory, sensors, batteries and radios mean that the price has dropped to, well, under a dollar in some cases. That makes it easy to lose track of such devices -- and forget that anything connected to our corporate network, or to the Internet, is both at risk of a breach and a potential threat if it's compromised, and then subverted to hurt us or others, electronically or in the real world.
Here are five suggestions for enterprises that are currently using embedded or IoT devices (which is probably most of them) or are considering (which is most of the rest). Anything connected to a network with a microprocessor is an IoT device, whether you think of it that way or not. That means risks. Security is not an option -- it's totally required.
1. Look for security information from device makers
Before you attach anything --anythingto your network (or allow a remote connection), make sure you understand its security basics. What are the built-in security features, such as the ROMs (read-only memory) it boots up from? What protections are built into its operation system and microprocessors? Where do patches come from, and how are they delivered to you (or the device) by the vendor?
More questions: Can you remotely manage the device or kill it if necessary, or are those things the vendor can do if you can't? If there is a breach or a reported vulnerability, what channels will the vendor take to notify you? How fast will the vendor promise to respond? Is there are service-level guarantee on how fast patches will be delivered? Lots and lots of questions. You need answers. If you don't like the answers, don't install the device -- because as soon as you do, you've assumed all the risk.
2. Install strong automated asset management
It's 2:00 a.m. Do you know where all your IoT devices are? An organization with large-scale deployments of mobile and portable devices must keep track of them all, but it's impossible to do so manually. The only solution is an automated asset management system that will listen for traffic for occasionally connected devices and reach out to contact devices on a regular schedule. (The schedule will vary depending on the device, of course.)
The asset management system and its agents should be able to monitor report on the health of the devices, check the ROMs and installed software to ensure they are up-to-date and not tampered with. If the device can provide its geolocation, that should be monitored as well, and alerts sent if something is wrong -- think geofencing. (Location can be approximated based on network connectivity.) If possible, the asset management system should be able to temporarily or permanently kill an IoT device, in the event of a breach, theft, or who-knows-what.
That's only the start. Data goes both ways. Make sure that data flowing from the device to your data center, or to a cloud service, can only be received by the correct recipient(s). Also, make sure that only data from authorized sources can be sent to the IoT device. Consider using VPNs, if possible, to ensure that the device is on a secure network. That's especially true if, for example, you're dealing with a portable or mobile device that's connected via WiFi. And encryption -- well, presumably everyone understands why nothing should go via clear text, and why encryption should be strong, up to date and not use buggy protocols or libraries. Right?
4. Ensure devices can be upgraded over the air
Not long ago, software updates to embedded devices required physical access to the device, to swap out a ROM, plug in an RS-232 patch cable, or place the device into "update" mode. Nowadays, forget that. Devices need to be updated "over the air" (we use that phrase even if the device is physically connected to the network), or they're not going to be updated at all.
There are questions. Who pushes out the updates, you or the vendor? It's best if you can, so that you can test updates and schedule the push. (Sometimes updates change security settings, which is one reason why you need to test.) There should be provisions to ensure that only authorized users (or authorized servers) can push new software into devices, whether it's firmware, operating systems, or applications residing in memory.
In the case of occasionally connected devices, be sure to determine whether you want them to update as soon as they connect, or if that's inconvenient for the user. Make sure you are in charge of updates, and that your asset management system knows what has been updated and where updates may have failed. It happens.
5. Remove insecure and orphaned devices immediately
Devices get old. Support contracts and bug-fixes can be discontinued by the manufacturer. Sometimes there's a show-stopped security flaw that simply can't be remediated through firmware. Make sure that you know about these situations -- and don't leave those devices active any longer than necessary. If you can swap out the hardware quickly, great. If not, consider deactivating the devices, or at least requiring the users to turn in the equipment.
All of this can be very tricky in the IoT domain -- imaging medical devices, farming, remote telemetry and so on. It's a lot harder than scheduling to replace laptops every three or five years! Getting the new devices and then scheduling and performing the device exchange can take a lot of time and cost a lot of money. That's a shame, but consider the risk of leaving insecure and unsupported devices in the field, or on your network.
To emphasize: You need the ability to wipe and shut down devices if they are breached, stolen, lost, or whatever. Make sure that's a capability of your asset management system, and of the IoT devices you purchase and deploy.
Get ready to act
There's risk in deploying any technology on your network. Those risks multiply with the IoT, in large part because of the sheer numbers of devices and the fact that many are portable or mobile. It's often impossible to manage them the same way you'd manage a desktop or a server.
That risk is part of the IoT equation. Make sure you are in change of the security of those IoT endpoints -- because if you're not, someone else might be.
— Alan Zeichick is principal analyst at Camden Associates, a technology consultancy in Phoenix, Arizona, specializing in enterprise networking, cybersecurity and software development. Follow him @zeichick.