The present application claims the benefit under 35 U.S.C. § 119 of German Patent Application No. DE 102015224868.2 filed on Dec. 10, 2015, which is expressly incorporated herein by reference in its entirety.
The present invention relates to an embedded system secured against attacks.
An electronic computer that is incorporated (“embedded”) in a technical working environment is commonly referred to in information technology as an embedded system. Here, the computer typically performs monitoring, control, or regulating functions, or is responsible for a form of the data or signal processing.
Conventionally, embedded systems are used in a large number of areas of application and equipment. In the case of complex overall systems, for example so-called sensor networks, this sometimes includes a large number of individual systems that for example are networked wirelessly.
German Patent Application No. DE 10 2013 213314 (A1) describes a method for storing at least one calculable integrity measurement value in a first memory area of a memory of a component, the at least one calculable integrity measurement value being ascertained as a test value of a software module of the component, and stored in the first memory area by an invariable write unit. In this way, a balance is sought between safety requirements and saving time during a booting or start process.
The present invention provides an embedded system secured against attacks.
An advantage of this solution is its improved protection against harmful software and system crashes. To this extent, a system according to the present invention is able to determine whether it is currently being attacked, or whether its software is displaying trustworthy, expected, and normal behavior.
The proposed approach here is based on the basic idea of analyzing the power consumption of one or more components in a system in order to recognize an untrustworthy behavior. The corresponding analysis component can be a hardware module installed in the system, designated in the following as “watchdog.” Countermeasures can be carried out when an unusual power signature is recognized in order to bring the system back to a trustworthy state.
In this way it can be recognized whether hardware components have really been used, for example whether a physical random number generator or a structure provided for the secure production or storing of secret keys (physical unclonable function) was really used to obtain a highly secure random number or, in the latter case, an authentic response.
Energy-consuming attacks on embedded devices having a limited battery capacity are also easy to recognize according to the present invention. A countermeasure can be to put the device into sleep mode.
If malicious software is recognized, e.g., if packets are sent outside the standard time window—sensor nodes typically send approximately one message per hour—then the system can be reset by reloading a trustworthy firmware from the read-only memory (ROM).
A system crash can be detected if the microcontroller, in contrast to all the other components, has a high energy consumption, but only when this behavior is not expected over a defined time interval. In this case, the watchdog can initiate a restart of the system.
Current peaks, for example of a radio interface, can indicate an active attack. The same conclusion is indicated if the network traffic within a business is monitored and too much data is received or sent.
Malware, active attacks, or system crashes can be acquired via a power signature model. As a function of the capacities of the device, this model can be very economical or can make higher demands with regard to its memory requirements.
Side channel properties, as normally used in the context of so-called side channel attacks—such as a power consumption profile of the main processor (central processing unit, CPU), memory side attacks and errors, access to the registers, etc.—are used to confirm normal behavior of the software.
The watchdog can be implemented in hardware or in at least one module that cannot be modified by a possibly malicious computer. To this extent, it represents the manipulation-proof root (root of trust) of a security chain (chain of trust) in the sense of the design known as “trusted computing” (TC).
Through the measures described herein, advantageous developments and improvements of the present invention are possible.
Exemplary embodiments of the present invention are shown in the figures and are explained in more detail below.
An embedded system (10) that includes a power supply (13) of the system (10), a computer (11), and the watchdog (12) acting according to the present invention, is shown in
Thus, the watchdog (12) constantly performs a comparison (17) of the first power signature (15), ascertained on the basis of the currently measured power consumption (14), with an expected second power signature (16). In the simplest case, this can take place using a mean value of the power consumption (14), or according to somewhat more sophisticated models. Because, with regard to the power consumption (14), the sending and receiving of WLAN data is more costly than purely local calculations, too much data traffic, or no data traffic at all, can be reliably recognized even on the basis of a mean value of the power consumption (14).
The computer (21) according to
In the scenario in
Number | Date | Country | Kind |
---|---|---|---|
10 2015 224 868 | Dec 2015 | DE | national |
Number | Name | Date | Kind |
---|---|---|---|
5341503 | Gladstein | Aug 1994 | A |
7612656 | Koie | Nov 2009 | B2 |
7877621 | Jacoby | Jan 2011 | B2 |
8332945 | Kim | Dec 2012 | B2 |
9268938 | Aguayo Gonzalez | Feb 2016 | B1 |
20050144612 | Wang | Jun 2005 | A1 |
20070006306 | Seifert | Jan 2007 | A1 |
20100168727 | Hancock | Jul 2010 | A1 |
20120186691 | Bridges | Jul 2012 | A1 |
20140108843 | Choho | Apr 2014 | A1 |
20140143597 | Chen | May 2014 | A1 |
20140223554 | Roden, III | Aug 2014 | A1 |
20150254017 | Soja | Sep 2015 | A1 |
20160232010 | Dicks | Aug 2016 | A1 |
Number | Date | Country |
---|---|---|
102013213314 | Jan 2015 | DE |
Number | Date | Country | |
---|---|---|---|
20170169221 A1 | Jun 2017 | US |