Each minute of every day 204 million emails are sent, 4 million Google searches are made, 277.000 tweets are published. This amazing amount of data is still mostly coming from people interacting with the digital world which in a few years has become the source of data driven technology with new business, services and revenue models. But with the IoT and its promises now ready to spread all over the world, an even bigger ocean of data will soon come directly from many sources (devices) to a lot of “things” (even more devices)! This new era will shape the embedded computer industry as much as the internet itself did a few years back. Let’s see why and how defense computing will also be transformed along the way.
The main business promises of IoT has focused around cost saving (eg: preventive maintenance) and new revenue streams (eg: specific services created from the collection of data from a multitude of device sensors) . To support that, the technologies and solutions being designed in this sector will simplify and standardize technical interoperability and business processes as well as boost business growth by creating new revenue streams (just imagine marine weather gathered by large carrier companies from each of their cargo vessels or detailed temperature, vibration sensors in a container allowing precious cargo to be individually tracked). From a high level point of view, the main pillars of IoT are: 1. collect (locate, secure), 2. analyze (authenticate, decrypt, aggregate), 3. manage (discover, commission, decommission), 4. monetize (sell, ship); IoT’s superior goal being to transform object data into actionable information and monetization. To achieve this, interoperable data stacks will replace data silos which are quite the norm in industrial computing today. The IoT will connect the power of cloud analytics with endpoints and remove current barriers. And for defense computing, IoT will propose solutions for failure analysis, device intelligence, measurement verification, anomaly and errors detection, actionable intelligence, capacity planning and customer intelligence. All topics which are already part of large defense systems of systems but are currently addressed ad hoc or with proprietary approaches.
Is the Embedded Computing Industry at a Turning Point?
I have always been shocked by the notion of “embedded PC” which has been the core of the largest segment of embedded computing in the last decade. How a computer can be personal AND embedded at the same time? The personal computer philosophy is to be friendly to any change – even if this change is driven from the outside (new downloadable software, latest gizmo plugged into a computer port). While on the other hand, embedded computing needs to focus on the highest computer integrity and continuous service. This is why we are left with this crazy situation where all the nice features of a personal computer become nasty issues for embedded computing. Current approaches to patch these inherent flaws today heavily rely on physical computer and network protections. As such, it is easy to understand that the overall connectivity and accessibility of all computers involved in the Internet of Things cannot be “embedded PCs”. The Internet of Things cannot deploy on PCs!
Yet, most of the offering today is still made of Box PCs, Blade PCs, or PC servers. So where is the solution? Can these challenges be solved quickly? Can a computer without any physical security fence be protected and act as a high integrity computing element while being exposed to all sorts of hacking? The answer is yes. There are solutions already available and they have been exposed to large scale real life scenario with great success. Where? Look at the game industry. In this segment, high end computers are deployed worldwide and are consider secure without physical protection. Because of this industry business model, the computer integrity must be guaranteed at all costs (not quite, since the unit cost is also a challenge here). Even when one has complete access to these devices, they cannot risk the integrity of the device, which would lead to a huge loss of revenue for this industry. Last time I checked, the PS4, deployed in November 2013 has not been compromised, (even though a method to pirate and reproduce game licenses is being prosecuted by Sony in Brazil). Two years in the hands of the most active hacking community is quite a record!
So what can we expect from the silicon industry and embedded computer vendors? Where can we expect more innovation? Where are we going after a long decade of PC dominance? And how will this influence defense computing?
The Internet of Things and Defense Computing: The IoT will Boost Trusted Operational Computing
Within the last decades, we have run through many eras of technical innovation. First, the microprocessor era, which led us from mainframes to minis and workstations. Then the PC era of the last 15 years when the PC architecture deployed in huge numbers albeit in various skins. Now we live in the object and cloud era with « all connected » billions of devices. And these devices will no longer be a personal computer in disguise. They will have to embark the best of breed of computer and software protections to be eligible for the Internet of Things.
For the embedded computing industry there is an urgent need for trusted computing. Not only for defense computing but for all kind of more common applications (think about your home alarm system), the issues the IoT infrastructure has to solve are: Is the data trustworthy? Is the data really coming from this computer? Is this Computer genuine? Is this computer uncompromised (since fabrication, or since commissioning)? Is the running software authentic? Is this computer operating properly?
It’s no question that the IOT will boost the solutions for trust in operational computers. And the embedded computer technology (ECT) will have to support the infrastructure for this.
Trust in a computer actually relies on a lot of parameters and a truly trusted operational system must be built from the ground up with security in mind. And for this, my definition of trust is a lot wider than just including a TPM chip. Kontron is already working on trusted operational computing, which encompasses all aspects of securing the data coming from a system. This covers health management information as well as tools which can also be used by non-experts and 3rd party software. The real job of an embedded computer vendor will no longer stop at the factory dock. Key ingredients to security such as digitally signing each computer during the manufacturing process, providing a trusted low level software baseline (where all the code executed is measured from the first instruction after reset), remote authentication, continuous comparison of the computer configuration against a reference profile and IT and OT payload segregation will become part of the COTS offering.
Being aware of these trends and solutions, Defense computing can benefit from them right away instead of creating everything from scratch. Why would this very capable industry turn away from top secret sauce and proprietary solutions only? Because of the necessary connectivity with the real world! Since the full scale deployment of the IoT concept will also contribute to digitizing and automating a bigger part of our world, even the most secure weapon systems will have to include more and more data from smart objects and their associated technology. This digital world will offer a whole new source of data points through key sensors, which in turn will foster innovation in weaponry and military counter measures. Leveraging technologies boosted by the Internet of Things has to be managed as an important disruptive step in sensor fusion, leading the way to even more sophisticated situational awareness applications.
To prepare for this, selecting the right technologies and suppliers is the big challenge for every defense contractor architect of today.
What is your opinion on trusted operational computing? Are you already designing for it?
Your comment was submitted.
An error occured on subscribing!: