US20180048550A1 - Device fingerprinting for cyber-physical systems - Google Patents

Device fingerprinting for cyber-physical systems Download PDF

Info

Publication number
US20180048550A1
US20180048550A1 US15/556,136 US201615556136A US2018048550A1 US 20180048550 A1 US20180048550 A1 US 20180048550A1 US 201615556136 A US201615556136 A US 201615556136A US 2018048550 A1 US2018048550 A1 US 2018048550A1
Authority
US
United States
Prior art keywords
network
fingerprinting
time
devices
physical
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US15/556,136
Other languages
English (en)
Inventor
Abdul Raheem Beyah
David Formby, Iii
Preethi Srinivasan
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Georgia Tech Research Corp
Original Assignee
Georgia Tech Research Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Georgia Tech Research Corp filed Critical Georgia Tech Research Corp
Priority to US15/556,136 priority Critical patent/US20180048550A1/en
Assigned to GEORGIA TECH RESEARCH CORPORATION reassignment GEORGIA TECH RESEARCH CORPORATION ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: BEYAH, ABDUL RAHEEM, FORMBY, David, SRINIVASAN, Preethi
Publication of US20180048550A1 publication Critical patent/US20180048550A1/en
Assigned to NATIONAL SCIENCE FOUNDATION reassignment NATIONAL SCIENCE FOUNDATION CONFIRMATORY LICENSE (SEE DOCUMENT FOR DETAILS). Assignors: GEORGIA INSTITUTE OF TECHNOLOGY
Abandoned legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
    • H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0876—Network utilisation, e.g. volume of load or congestion level
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F15/00—Digital computers in general; Data processing equipment in general
    • G06F15/16—Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/06—Generation of reports
    • H04L43/065—Generation of reports related to network devices
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
    • H04L63/0876—Network architectures or network communication protocols for network security for authentication of entities based on the identity of the terminal or configuration, e.g. MAC address, hardware or software configuration or device fingerprint

Definitions

  • Fingerprinting devices on a target network can provide network administrators with mechanisms for intrusion detection or enable adversaries to conduct surveillance in preparation for a more sophisticated attack.
  • ICS industrial control systems
  • An attacker intruding on a network can theoretically inject false data or commands and drive the system into an unsafe state.
  • Example consequences of such an intrusion can range from widespread blackouts in a power grid to environmental disasters caused by tampering with systems carrying water, sewage oil, or natural gas. These false data and command injections could be thwarted using strong cryptographic protocols that provide integrity and authentication guarantees.
  • FIG. 1 is a schematic representation illustrating an example of a configuration for fingerprinting performed in a control system environment according to various embodiments of the present disclosure.
  • FIG. 2 is a diagram illustrating a measurement of a cross-layer response time according to various embodiments of the present disclosure.
  • FIG. 3 is a schematic representation illustrating an example of different points of attack that an adversary may exploit when attacking power substation network.
  • FIGS. 4 a and 4 b show examples of network architectures used to test a cross-layer response time method of device fingerprinting according to various embodiments of the present disclosure.
  • FIGS. 5 a and 5 b show an example scatterplot of cross-layer response times for sample ICS devices and corresponding probability density functions (PDFs) for the sample ICS devices.
  • PDFs probability density functions
  • FIG. 6 shows an example of fingerprint classification performance using FF-ANN.
  • FIGS. 7 a and 7 b show estimated probability density functions (PDFs) for the sample ICS devices after upgrades to network architecture and an increase in polling frequency.
  • PDFs estimated probability density functions
  • FIG. 8 is a flowchart illustrating one example of a method of cross-layer response time device fingerprinting according to various embodiments of the present disclosure.
  • FIG. 9 is a timing diagram illustrating a calculation physical operation times according to various embodiments of the present disclosure.
  • FIG. 10 is a schematic representation illustrating an example of a configuration for testing physical device fingerprinting according to various embodiments of the present disclosure.
  • FIG. 11 shows graphs illustrating the distribution of close operation times based on SER responses and open operation times based on SER responses.
  • FIG. 12 is a flowchart illustrating one example of a method of physical device fingerprinting according to various embodiments of the present disclosure.
  • FIG. 13 is a schematic block diagram that provides one example illustration of a computing environment employed in the control system environment of FIG. 1 , according to various embodiments of the present disclosure.
  • Embodiments of the present disclosure provide for device fingerprinting in cyber-physical system, such as a control system environment.
  • Embodiments of the present disclosure can be used in conjunction with traditional intrusion detection system (IDS) in a control systems environment.
  • IDS intrusion detection system
  • Embodiments of the present disclosure can be used to achieve device fingerprinting from software, hardware, and physics-based perspectives.
  • Embodiments of the present disclosure can prevent security compromises by accurately fingerprinting devices in a control system environment, and other networked environments, as may be appreciated.
  • Embodiments of the present disclosure can generate fingerprints of a device which reflects identifiable characteristics of a device, such as, e.g., processing speed, processing load, memory speed, and protocol stack implementation.
  • a network monitoring device can constantly monitor all traffic on a network.
  • the network monitoring device can be installed in a communication path.
  • the network monitoring device can listen to a port that mirrors all traffic on the network.
  • the network monitoring device can be a tap.
  • a master device can send read requests for measurements over the network to field devices operating in a control systems environment.
  • the field devices can send responses in return.
  • the network monitoring device can parse fields in the network traffic at a transmission control protocol (TCP) level and a control system application layer.
  • TCP transmission control protocol
  • the network monitoring device can parse application layer headers.
  • the network monitoring device can store identifying information for each of the read requests.
  • the network monitoring device can record times when a TCP acknowledgment (ACK) is seen for each of the read requests.
  • TCP transmission control protocol
  • the network monitoring device can store a time when each response appears for every read request.
  • the network monitoring device can measure an amount of time between the TCP ACK and the time when each response appears for every read request, referred to as a cross-layer response time (CLRT).
  • CLRT cross-layer response time
  • a fingerprint for each field device can be generated based at least in part upon the amount of time between the TCP ACK of each of the read requests and the appearance of each corresponding response.
  • the fingerprint can be represented as a probability density function (PDF) of the measured amounts of time between the TCP ACK and the time when each response appears for every read request.
  • PDF probability density function
  • a minimum threshold number of response times can be calculated before a fingerprint can be generated.
  • a network monitoring device can constantly monitor all traffic on a network.
  • the network monitoring device can be installed in a communication path or can listen to a port that mirrors all traffic on the network.
  • the network monitoring device can be a tap.
  • the network monitoring device can be a sniffer used to parse packets to perform deep packet inspection.
  • a master device can send a command to a field device to perform a task or an operation.
  • a slave device can be hardwired to the field device.
  • a slave device can be connected to the field device via a digital network (e.g., Ethernet). Responses to the command from the field device can be observed at the slave device.
  • the slave device can asynchronously respond to the master device with a message indicating an event change.
  • the event change can be observed with a network tap to calculate an operation time of the field device in responding to the command.
  • an unsolicited response timestamp can be calculated at the tap point by measuring the difference between a time at which the command was observed and a time at which the response was observed to get a measurement of physical device response time.
  • the physical field device operation times can be calculated by and stored in the slave device and later transmitted to the master.
  • a sequence of event recorder response time can be calculated by measuring the difference between a time at which was the command was observed at the tap point and an event timestamp performed by an application layer.
  • a fingerprint can be generated based at least in part upon the unsolicited response time. In other embodiments, a fingerprint can be generated based at least in part upon the sequence of event recorder response time. In some embodiments, a minimum threshold number of response times can be calculated before a fingerprint can be generated.
  • the present disclosure is not limited to particular materials, reagents, reaction materials, manufacturing processes, dimensions, frequency ranges, applications, or the like, as such can vary. It is also to be understood that the terminology used herein is for purposes of describing particular embodiments only, and is not intended to be limiting. It is also possible in the present disclosure that steps can be executed in different sequence, where this is logically possible. It is also possible that the embodiments of the present disclosure can be applied to additional embodiments involving measurements beyond the examples described herein, which are not intended to be limiting. It is furthermore possible that the embodiments of the present disclosure can be combined or integrated with other measurement techniques beyond the examples described herein, which are not intended to be limiting.
  • the fingerprint (or signature) of a device can be represented as a probability density function (PDF) of the response times of devices in a cyber-physical system.
  • PDF probability density function
  • white box a dynamic model of the device is constructed from principles and model parameters identified from CAD drawings, source code, physical measurements, etc. without ever seeing any true samples from the system.
  • the simulated behavior is then used to create a PDF by varying model parameters using an uncertainty distribution.
  • black box approach the PDF is constructed strictly from experimental data without any dynamic modeling. Black box modeling requires a significant amount of experimental measurements, but little knowledge of the underlying system.
  • gray box modeling is best suited for when a system's internal details are accessible, but access to experimental measurements is restricted. Black box modeling performs best when experimental measurements are easily available, and is especially effective when the system is proprietary or too complex to model.
  • gray box modeling approaches are most advantageous when the basic characteristics of a software or hardware design are known, but there is some uncertainty in model structure or parameters that can only be dealt with through experimental observations.
  • cross-layer fingerprinting Due to the abundance of measurements in the available dataset and lack of proprietary source code, the data acquisition fingerprinting method called cross-layer fingerprinting, focuses on a black box modeling approach.
  • the physical fingerprinting technique there are some devices where the operations occur so rarely that collecting enough real samples to generate an accurate fingerprint through black box modeling can be completely infeasible. Additionally, there is such a wide variety of physical devices available and their costs are so prohibitive that creating a black box signature database offline is also infeasible. Therefore an alternative approach for signature generation can be used.
  • a new class of fingerprint generation for physical fingerprinting based on white box modeling allows an administrator to generate a usable device fingerprint without ever having access to the target device type or network.
  • the white box-generated physical fingerprint is then validated against the black box approach using an example control device.
  • the approaches described herein take advantage of the unique characteristics of ICS devices and other control systems devices.
  • a new class of fingerprint generation specific to ICS networks using “white box” modeling is shown.
  • the various embodiments of the present disclosure also show performance analysis using both real world data from a power substation and controlled lab tests.
  • the methods of fingerprint generation according to various embodiments of the present disclosure can be evaluated under simple forgery attacks for different classes of adversary.
  • Device fingerprinting methods are usually classified into active or passive techniques depending on whether they actively probe a device with specially crafted packets or passively monitor network traffic to develop the fingerprint.
  • One of the oldest fingerprinting tools, Nmap® uses active fingerprinting techniques to gather information about devices on a network. By sending a series of specific requests, Nmap® determines the operating system (OS) and server versions running on a machine based on how the device responds. While this tool is invaluable for both pen-testers and attackers on a “normal” network, it has limited use in an ICS network where active methods are not as desirable.
  • OS operating system
  • ICS For passive fingerprinting, a variety of techniques exist that provide both device type fingerprinting and individual device fingerprinting.
  • One example is the open source p0f tool, which passively examines TCP and hypertext transfer protocol (HTTP) header fields to determine information about a client, such as OS and browser version.
  • HTTP hypertext transfer protocol
  • the first attempt at formalizing methods for active and passive fingerprinting of network protocols was published in 2006, when parametrized extended finite state machine (PEFSMs) were used to model the behavior of different protocol implementations. See G. Shu and D. Lee. Network protocol system fingerprinting—a formal approach. In INFOCOM 2006. 25 th IEEE International Conference on Computer Communications. Proceedings , pages 1-12, April 2006. Determining software versions is of some use, but identifying individual devices on a network based on their hardware is even more useful, which for example, could be used for tracking a device across the Internet or intrusion detection.
  • PEFSMs parametrized extended finite state machine
  • Ptf Passive temporal fingerprinting. In Integrated Network Management (IM ), 2011 IFIP/IEEE International Symposium on, pages 289-296, May 2011. A third paper that used passive observations of network traffic timing to achieve device fingerprinting was published in 2014, and used distributions of packet inter-arrival times (IAT) to identify devices and device types. See S. Radhakrishnan, A. Uluagac, and R. Beyah. Gtid: A technique for physical device and device type fingerprinting. Dependable and Secure Computing, IEEE Transactions on, PP(99):1-1, 2014.
  • Another approach to passive device fingerprinting focuses on the physical layer of device communication, rather than the higher layers. Specifically, amplitude and phase measurements of the signals generated by Wi-Fi radios were used to identify individual devices. However, using amplitude and phase measurements of the signals generated by Wi-Fi radios is still is not feasible in ICS networks, and other control systems networks where Wi-Fi devices are rarely used.
  • the fingerprinting techniques overcome the limitations of previous works on device fingerprinting by providing higher accuracy results using techniques that are especially suited for ICS and other cyber-physical systems.
  • One embodiment of the present disclosure improves on more traditional timing-based approaches by using network traffic measurements that are unique to control systems devices.
  • the idea of physical layer fingerprinting is extended to identifying ICS control devices based on the reported timings of each device's physical operations.
  • all previous fingerprinting work used black box methods that require access to an example target device.
  • Various embodiments of the present disclosure overcome this limitation by proposing a white box fingerprint generation approach that does not need previous access to example devices.
  • the first attempt at tailoring IDS methods for ICS and supervisory control and data acquisition (SCADA) systems focused on monitoring traffic flows for regular patterns and understanding packets at the application layer to look for intrusions.
  • Some researchers have also approached the problem by modifying IDS software to perform specification based intrusion detection for common ICS protocols.
  • Others have attempted to model the states that a process control system can enter and detect when a command might cause it to enter a critical state. These solutions are able to detect some types of attacks, but are unable to detect a class of stealthier ones called false data injection attacks.
  • the first attacker model was chosen due to how vulnerable these devices are (as evidenced by the 30 year old TCP vulnerabilities found widespread in the power grid) and because it was the method used on the most well-known ICS attack to date, Stuxnet.
  • the second attacker model is realistic in the scenario of a widely distributed control system where physical security can be difficult to achieve.
  • FIG. 1 shown is a schematic representation illustrating an example of a configuration for performing methods of device fingerprinting in a control system environment 100 in conjunction with a traditional intrusion detection system (IDS) 103 .
  • IDS intrusion detection system
  • the methods of device fingerprinting discussed herein can achieve device fingerprinting from software, hardware, and physics-based perspectives.
  • a network monitoring device can be installed in a communication path 203 or can listen to a port that mirrors all traffic on the network.
  • Many control systems protocols operate on top of TCP and use a Read/Response architecture where a master station can send Read requests for measurements to devices in the field and the devices send Responses in return.
  • the network monitoring device can parse fields in network traffic at the TCP level and control system application layer.
  • the network monitoring device can parse the application layer headers, store identifying information for each Read request, record the times when the TCP ACK is seen for each Read request, and store the time when the corresponding Response appears for every Read request.
  • a CLRT measurement is the time between when the TCP layer acknowledges that the Read request packet was received and when the application layer sends the Response.
  • a CLRT measurement can be obtained by directly measuring the application layer Response time.
  • a fingerprint is generated for each device 206 based on the distribution of time that the device 206 takes to process the request.
  • the timing diagram 200 shows how a CLRT measurement can be taken in a typical SCADA network or any other control network, as may be appreciated. It should be noted that since the CLRT measurement is based on the time between two consecutive packets from the same source to the same destination, the measurement can be independent of the round trip time between the two nodes.
  • the fingerprint signature can be defined by a vector of bin counts from a histogram of CLRTs where the final bin includes all values greater than a heuristic threshold.
  • M be a set of CLRT measurements from a specific device
  • B define the number of bins in the histogram (and equivalently the number of features in the signature vector)
  • H signify the heuristic threshold chosen to be an estimate of the global maximum that CLRT measurements should ever take.
  • the range of possible values can be divided by thresholds t i where
  • the CLRT measurement is advantageous for fingerprinting ICS devices because it remains relatively static and its distribution is unique within device types and even software configurations. To understand why this is true for ICS devices, all of the factors which might affect this measurement must be considered.
  • ICS devices can have simpler hardware and software architectures than general purpose computers because ICS devices are built to perform very specialized critical tasks and can do little else.
  • a typical modern-day computer now has fast multi-core processors in the range of 2-3 GHz with significant caching, gigabytes of RAM, and context switching between the wide variety of processes running on the machine.
  • the ICS world is dominated by PLCs running on low powered CPUs in the tens to hundreds of MHz frequencies with little to no caching, tens to hundreds of megabytes of RAM, and very few processes. With such limited computing power available, relatively small changes in programming result in observable timing differences.
  • ICS device types are built with different hardware specifications (CPU frequencies, memory and bus speeds) as well as different software (operating systems, protocol stack implementations, number of measurements being taken, complexity of control logic) all resulting in each one being able to process requests at different speeds.
  • CPU frequencies, memory and bus speeds CPU frequencies, memory and bus speeds
  • software operating systems, protocol stack implementations, number of measurements being taken, complexity of control logic
  • CLRT measurements can be leveraged to identify ICS device types, but this does not explain why the CLRT measurements are so constant over the network.
  • FIG. 3 shown is a schematic representation that illustrates different points of attack that an adversary can exploit when attacking a power substation network 300 .
  • the adversary 303 can either attack a communication infrastructure or one of a number of individual devices such as a remote terminal unit (RTU) 306 or a programmable logic controller (PLC) 309 .
  • RTU remote terminal unit
  • PLC programmable logic controller
  • the adversary 303 can attempt to inject false data responses, false command responses, or both.
  • false data and command injections such as these can have disastrous effects on a power grid. Therefore, the device fingerprinting techniques described herein can identify what type of devices these responses are originating from.
  • Such fingerprinting techniques can be important to distinguishing between responses originating from a legitimate intelligent electronic device (TED), an adversary 203 with a laptop who has gained access to the network 300 , and a comprised TED posing as a different device on the network 300 .
  • FIGS. 4 a and 4 b shown are examples of network architectures 400 a and 400 b used to test a cross-layer response time (CLRT) method of device fingerprinting.
  • CLRT cross-layer response time
  • the CLRT measurements are much larger than most delays that might be caused by differences in network architecture.
  • the CLRT measurements are all on the order of tens or even hundreds of milliseconds.
  • typical latencies obtained from ICS network switch datasheets and theoretical transmission delays on a 100 Mbps link are both on the order of microseconds, resulting in a minor contribution to the overall CLRT measurement.
  • ICS networks most often have overprovisioned available bandwidth to ensure reliability (e.g.
  • the first substation network studied for this research used an average of 11 Kbps bandwidth out of the available 100 Mbps, a strikingly low traffic intensity of 0.01%). These low traffic intensities ensure that the switches and routers on the network are never heavily loaded and have consistently low queuing delays.
  • FIG. 4( a ) shows a first network architecture 400 a of a first substation 403 a.
  • Approximately 20 GB of network traffic was captured from the first substation 403 a with approximately 130 field devices running a distributed network protocol (DNP3) over a span of five months within the first network architecture 400 a.
  • DNP3 distributed network protocol
  • one more month of data was captured from the first substation 403 a after the first network architecture 400 a was slightly modified by replacing the main router with a new switch, changing the IP addressing scheme accordingly, and increasing the frequency of measurement polling.
  • DNP3 distributed network protocol
  • the second architecture 400 b comprises approximately 80 field devices using DNP3 to test if device fingerprints learned on the first network architecture 400 a would translate to another network architecture. Further tests were conducted in the lab to study the effects on the software configuration alone and to rule out any possible factors related to different hardware or different round-trip times (RTT) on the network.
  • RTT round-trip times
  • cross-layer response time measurements were taken from DNP3 polling requests for event data and were summarized by dividing all measurements into time slices (e.g., one hour, or one day) and calculating means, variances, and 200-bin histograms for each time slice.
  • Machine learning techniques were then evaluated using two different feature vectors: a more complex approach using the arrays of bin counts as defined in the equation below and a simple approach using arrays containing only the mean and variance for each time slice.
  • FIG. 5( a ) shown is an example of a scatterplot of cross-layer response times (CLRTs) for sample ICS devices.
  • the corresponding probability density functions (PDFs) 500 b for the sample ICS devices are shown in FIG. 5( b ) .
  • PDFs 500 b probability density functions
  • FIG. 5( b ) To obtain a rough visualization of the separability of the device types based on their CLRT measurements, a scatter plot 500 a based on the mean and variances of CLRTs was produced and the true labels of the devices are illustrated in FIG. 5( a ) .
  • Each point in the scatter plot 500 a represents the mean and variance of the CLRT measurements for one IP address over the course of one day out of the original five month dataset.
  • the scatter plot 500 a shows the highly separable hardware device types of Vendor A (Types 1a 503 , 1b 506 , and 2 509 ), Vendor B 512 , and Vendor C 515 . Furthermore, the scatter plot 500 a shows that identical hardware device types can be subdivided into classes based on different software configurations represented by Vendor A Type 1a 503 and Vendor A Type 1b 506 . These conclusions are further supported when the corresponding PDFs 500 b of CLRTs over a day were estimated for each type in FIG. 5( b ) .
  • FIG. 5( a ) illustrates that device types are clearly separable based on simple mean and variance measurements, many choices of a properly-tuned machine learning algorithm can result in high accuracy classification. Therefore, a sampling of the most popular algorithms in the field were chosen as examples.
  • the standard classification metrics of accuracy, precision, and recall as defined in Equations 2, 3, and 4 are calculated for each class separately, where TP, TN, FP, and FN stand for true positive, true negative, false positive, and false negative, respectively. To summarize these results, the average value across classes was plotted alongside the minimum value among classes.
  • FIG. 6 shown are results of fingerprinting classification performance using a feed forward artificial neural network (FF-ANN) with one hidden layer trained using the back propagation algorithm.
  • FF-ANN feed forward artificial neural network
  • FIG. 7( a ) shown is the CLRT distribution 700 a after changes were made to the first network architecture in FIG. 4( a ) .
  • the CLRT distribution with the new network architecture in FIG. 4( b ) is compared with the original in FIG. 4( a ) .
  • the fingerprints learned from the original capture were tested on the new data, very high accuracies in were obtained, which shows that the method is stable over long periods of time and over minor changes in the same network.
  • the primary defensive use-case for this technique would involve a training period on the target network.
  • an administrator is able to learn fingerprints on one network because of known labels, but does not have the labels for a different network is considered.
  • fingerprints from the original capture were studied and tested on a different substation over a year later.
  • the different substation's distribution in FIG. 7( b ) is compared with the original there are some small, but noticeable changes that could be result of the different architecture affecting the timings or from the different electrical circuit affecting the load of the devices.
  • the fingerprints learned from the original capture were tested on this different network, the average accuracy seemed to level off around 90%, suggesting that while the accuracy may be diminished across different networks, there is still some utility in the technique.
  • a read request is sent to a device in a control system.
  • the device may be for example, an RTU, an IED, or any other device in a control system, as can be appreciated.
  • a corresponding response to each read request is received from the device.
  • an amount of time between an acknowledgment of each read request and a time when the corresponding response is received is measured.
  • a fingerprint is generated for the device based on the amount of time that is measured.
  • a minimum number of measurements for the device can be required for the fingerprint to be generated. For example, a minimum threshold of 1000 samples (or other defined threshold) for the device may be required for the fingerprint.
  • a timing diagram 900 illustrating a calculation of physical operation times.
  • physical devices can be fingerprinted based on each device's unique physical properties and characteristics.
  • a series of operation time measurements can be taken and used to build an estimated distribution and generate a signature.
  • the formal definition of the signature follows the same logic as Equation 1 above, but with M being defined as a set of operation time measurements and H being a heuristic threshold chosen to be an estimate of the maximum value an operation should ever take.
  • the mechanical and physical properties defining how quickly a device operates differs between devices and produces a unique fingerprint for each device. For example, analyzing the difference in operation times of latching relays that use a solenoid coil arrangement shows that a unique fingerprint is produced for each device. Relays were chosen for this research as they are commonly used in ICS networks for controlling and switching higher power circuits with low power control signals.
  • the electromagnetic force produced while energizing the solenoid coil in a latching relay is directly proportional to current though the solenoid, number of turns in the solenoid, and the cross sectional area and type of core, as described by the equation below, where N is the number of turns in the solenoid, I is the current in amperes running through the solenoid, A is the cross-sectional area in meters-squared of the solenoidal magnet, g is the distance in meters between the magnet and piece of metal, and ⁇ 0 is the constant 4 ⁇ *10 ⁇ 7 .
  • FIG. 10 shown is a schematic representation illustrating an example of a configuration for testing physical device fingerprinting.
  • a circuit breaker operation was chosen.
  • the experimental setup consisted of a DNP3 master device 1003 from a C++ open source DNP3 implementation (OpenDNP3 version 2.0), an SEL-751A DNP3 slave device 1006 and two latching relays 1012 a and 1012 b to demonstrate physical device fingerprinting based on operation time.
  • the SEL-751A DNP3 slave device 1006 is connected to the two latching relays 1012 a and 1012 b by hardwired connections.
  • a C based DNP3 sniffer 1009 is used to sniff and parse the DNP3 packets to perform deep packet inspection.
  • the packets are timestamped by the Linux operating system which is time-synchronized by the same time source as that of the DNP3 master device 1003 and SEL-751A DNP3 slave device 1006 .
  • the SEL-751A IED is a feeder protection relay supporting Modbus, DNP3, IEC61850 protocol, time synchronization based on SNTP protocol, and a fast SER protocol which timestamps events with millisecond resolution.
  • the experimental setup for both relays 1012 a and 1012 b consisted of a latching circuit and a load circuit.
  • the latching circuit works on an operating voltage of 24 VDC needing about 1 A to operate and load circuit is based on 110V to be compatible with the LED's inputs.
  • the IED On a close command from the DNP3 master device 1003 , the IED activates a binary output energizing a latch coil to close the load circuit. Once the load circuit is energized, the binary input senses the change and a timestamped event is generated.
  • the IED activates the second binary output energizing the reset coil to open the load circuit, which is recorded as a timestamped event.
  • FIG. 11 shown are the distributions of close operation times based on SER timestamps for the latching relays 1012 a and 1012 b ( FIG. 10 ) from two different vendors.
  • the times range from 16 ms to 38 ms for Vendor 1 and 14 ms to 33 ms for Vendor 2.
  • both latching relays 1012 a and 1012 b FIG. 10
  • the difference in operation can be attributed to the difference in physical makeup between them.
  • one of the latching relays had a larger cross sectional area for its solenoid, resulting in different forces produced.
  • a command is sent from a master device to a field device to perform an operation.
  • the command can comprise, for example, a request to open a load circuit.
  • the command can comprise, for example, a request to close a load circuit.
  • an event change is observed at a slave device.
  • the slave device can be for example, a DNP3 slave.
  • the slave device can be connected to the field device via, for example, hardwire connections.
  • a message indicating the event change is sent by the slave device.
  • an operation time of the field device is calculated based at least in part upon a time at which the message was observed.
  • the message can be observed at, for example, a network tap.
  • a fingerprint for the field device is generated based at least in part upon the operation time. A minimum number threshold of measured operation times can be required before a fingerprint is generated.
  • the CLRT technique fingerprints and the physical device fingerprints were generated using black box methods that assume some access to the target devices.
  • the CLRT technique is based on monitoring of data packets requires a black box modeling approach as neither the internal circuitry nor the device source code is usually available (and thus there is no basis for constructing a white box model).
  • the physical device fingerprinting technique may leverage a white box, black box, or gray box modeling approach since the mechanical composition of a device can usually be obtained from manual inspection, available drawings/pictures, or manufacturer's specifications.
  • the ability to construct white box model fingerprints for physical device fingerprinting is important due to the rare operation of some devices, and the prohibitive cost of performing black box modeling on all of the available devices on the market.
  • construction of the same fingerprint for the latch relay mechanism is discussed using white box modeling only and then validates it against the black box model results obtained for the device.
  • a gray box modeling approach could be pursued as a general methodology for physical signature generation.
  • a standard latch relay operates using the principle of remanent magnetization in which a coil magnetizes a permanent magnet in either direction during opening and closing operations.
  • the latch relay was disassembled and its basic components modeled.
  • a magnetic armature of length L is connected to the base assembly by a torsional spring of spring constant k.
  • the torsional spring is preloaded so that it applies a torque which pushes the armature to the open position by default.
  • a permanent magnet lies at a distance l along the armature and is assumed to exert a magnetic force F p at a single point along the armature.
  • the permanent magnet is surrounded by a wire coil which carries the input current ⁇ (t), and also applies a magnetic force F c to the armature.
  • the magnetic field from the coil pulse drives the magnetic field of the permanent magnet to be in the same direction.
  • the permanent magnet holds the field in the same direction by the property of remanent magnetization. This process is what “latches” the relay.
  • I is the moment of inertia of the armature about the hinge point.
  • Physical measurements of the device can be used to provide values for r; R; l; L; k; and I.
  • Five other parameters must be identified to simulate the time response of the latch relay mechanism, namely c p ; c c ; ⁇ ; ⁇ , and ⁇ . These parameters may be estimated based on material composition of the magnets.
  • the white box model expectedly does not perform quite as well as the black box method based on true measurements due to the various simplifications and estimations made during the modeling process.
  • the results are still promising for this new class of fingerprinting.
  • the white box model approach would be limited to scenarios where there is not enough experimental data or the integrity of the experimental data is in question.
  • the white box approach can then be combined with the black box approach to enable gray box modeling where appropriate to achieve higher accuracy. While there are a variety of techniques to approach this problem, intuitively it is similar to simply replacing synthetic samples in the white box distribution with real samples over time as they become available.
  • the techniques should be relatively accurate and scalable.
  • Each method of device fingerprinting described herein achieved high enough accuracy for a defense-in-breadth strategy as a supplement to traditional IDS approaches.
  • the CLRT fingerprinting method achieved impressive classification accuracies as high as 99% in some cases and the physical device fingerprinting method was able to accurately classify measurements from two nearly identical devices around 92% of the time.
  • the FF-ANN algorithm used in training the two fingerprinting techniques only had one hidden layer and 200 input features, resulting in reasonable scalability for computational complexity.
  • the alternate Bayes classifier algorithm can also be efficient.
  • the results suggest that the accuracy for the methods scales as well.
  • the CLRT fingerprinting method was already tested above on a full scale power substation network and was able to achieve high accuracies.
  • the physical device fingerprinting method only achieved an accuracy of 92% for two similarly rated devices, an even higher accuracy can be expected as more diverse types of devices are added to the test set, resulting in more clear differences in distributions.
  • the fingerprints be nontrivial to forge (i.e., resistant to mimicry attacks).
  • the proposed methods of device fingerprinting are not so easily broken.
  • the compromised device since the ICS environment contains an abundance of legacy devices, it is not certain that the compromised device would even have a network card that supports promiscuous mode for network sniffing. Additionally, any sniffing code installed on a low powered, compromised device would most likely be computationally expensive enough to skew timing measurements on the system. Furthermore, since it was found that network architecture does have some effect on the fingerprint, this suggests that the adversary would have to sniff the network in the same location as the fingerprinter to get a completely accurate distribution, or be able to determine the effects of the network by other means.
  • the device fingerprinting techniques proposed here are passive and do not need changes to the target network or devices, better defenses against mimicry attacks could be implemented if this assumption is removed.
  • the SCADA master or the fingerprinter could be configured to randomly send extra requests or commands that have no effect on the operation of the network, but would increase the knowledge requirement of the adversary and the complexity of the behavior she has to mimic.
  • the CLRT method this could involve changing from polling for event data to polling for different numbers of specific measurements each time, which on the low powered embedded systems should theoretically result in measurable timing differences.
  • the physical fingerprinting method this could take the form of sending redundant commands, for example by sending a close command when the breaker is already closed.
  • the control system environment 100 includes one or more devices 1300 .
  • the device 1300 may represent a computing device or system (e.g. computers, servers, SCADA master, etc.).
  • Each device 1300 includes at least one processor circuit, for example, having a processor 1303 and a memory 1306 , both of which are coupled to a local interface 1309 .
  • each device 1300 may comprise, for example, at least one server computer or like device.
  • the local interface 1309 may comprise, for example, a data bus with an accompanying address/control bus or other bus structure as can be appreciated.
  • the device 1300 can include one or more network interfaces 1310 .
  • the network interface 1310 may comprise, for example, a wireless transmitter, a wireless transceiver, and a wireless receiver.
  • the network interface 1310 can communicate to a remote computing device using any of a variety of communication protocols as previously discussed. As one skilled in the art can appreciate, other communication protocols may be used in the various embodiments of the present disclosure.
  • Stored in the memory 1306 are both data and several components that are executable by the processor 1303 .
  • stored in the memory 1306 and executable by the processor 1303 are device fingerprinting program 1315 , application program 1318 , and potentially other applications.
  • Also stored in the memory 1306 may be a data store 1312 and other data.
  • an operating system may be stored in the memory 1306 and executable by the processor 1303 .
  • any one of a number of programming languages may be employed such as, for example, C, C++, C#, Objective C, Java®, JavaScript®, Perl, PHP, Visual Basic®, Python®, Ruby, Flash®, or other programming languages.
  • executable means a program file that is in a form that can ultimately be run by the processor 1303 .
  • Examples of executable programs may be, for example, a compiled program that can be translated into machine code in a format that can be loaded into a random access portion of the memory 1306 and run by the processor 1303 , source code that may be expressed in proper format such as object code that is capable of being loaded into a random access portion of the memory 1306 and executed by the processor 1303 , or source code that may be interpreted by another executable program to generate instructions in a random access portion of the memory 1306 to be executed by the processor 1303 , etc.
  • An executable program may be stored in any portion or component of the memory 1306 including, for example, random access memory (RAM), read-only memory (ROM), hard drive, solid-state drive, USB flash drive, memory card, optical disc such as compact disc (CD) or digital versatile disc (DVD), floppy disk, magnetic tape, or other memory components.
  • RAM random access memory
  • ROM read-only memory
  • hard drive solid-state drive
  • USB flash drive USB flash drive
  • memory card such as compact disc (CD) or digital versatile disc (DVD), floppy disk, magnetic tape, or other memory components.
  • CD compact disc
  • DVD digital versatile disc
  • the memory 1306 is defined herein as including both volatile and nonvolatile memory and data storage components. Volatile components are those that do not retain data values upon loss of power. Nonvolatile components are those that retain data upon a loss of power.
  • the memory 1306 may comprise, for example, random access memory (RAM), read-only memory (ROM), hard disk drives, solid-state drives, USB flash drives, memory cards accessed via a memory card reader, floppy disks accessed via an associated floppy disk drive, optical discs accessed via an optical disc drive, magnetic tapes accessed via an appropriate tape drive, and/or other memory components, or a combination of any two or more of these memory components.
  • the RAM may comprise, for example, static random access memory (SRAM), dynamic random access memory (DRAM), or magnetic random access memory (MRAM) and other such devices.
  • the ROM may comprise, for example, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or other like memory device.
  • the processor 1303 may represent multiple processors 1303 and/or multiple processor cores and the memory 1306 may represent multiple memories 1306 that operate in parallel processing circuits, respectively.
  • the local interface 1309 may be an appropriate network that facilitates communication between any two of the multiple processors 1303 , between any processor 1303 and any of the memories 1306 , or between any two of the memories 1306 , etc.
  • the local interface 1309 may comprise additional systems designed to coordinate this communication, including, for example, performing load balancing.
  • the processor 1303 may be of electrical or of some other available construction.
  • the device fingerprinting program 1315 and the application program 1318 may be embodied in software or code executed by general purpose hardware as discussed above, as an alternative the same may also be embodied in dedicated hardware or a combination of software/general purpose hardware and dedicated hardware. If embodied in dedicated hardware, each can be implemented as a circuit or state machine that employs any one of or a combination of a number of technologies. These technologies may include, but are not limited to, discrete logic circuits having logic gates for implementing various logic functions upon an application of one or more data signals, application specific integrated circuits (ASICs) having appropriate logic gates, field-programmable gate arrays (FPGAs), or other components, etc. Such technologies are generally well known by those skilled in the art and, consequently, are not described in detail herein.
  • each block may represent a module, segment, or portion of code that comprises program instructions to implement the specified logical function(s).
  • the program instructions may be embodied in the form of source code that comprises human-readable statements written in a programming language or machine code that comprises numerical instructions recognizable by a suitable execution system such as a processor 1303 in a computer system or other system.
  • the machine code may be converted from the source code, etc.
  • each block may represent a circuit or a number of interconnected circuits to implement the specified logical function(s).
  • FIGS. 8 and 12 show a specific order of execution, it is understood that the order of execution may differ from that which is depicted. For example, the order of execution of two or more blocks may be scrambled relative to the order shown. Also, two or more blocks shown in succession in FIGS. 8 and 12 may be executed concurrently or with partial concurrence. Further, in some embodiments, one or more of the blocks shown in FIGS. 8 and 12 may be skipped or omitted. In addition, any number of counters, state variables, warning semaphores, or messages might be added to the logical flow described herein, for purposes of enhanced utility, accounting, performance measurement, or providing troubleshooting aids, etc. It is understood that all such variations are within the scope of the present disclosure.
  • any logic or application described herein, including the device fingerprinting program 1315 and the application program 1318 , that comprises software or code can be embodied in any non-transitory computer-readable medium for use by or in connection with an instruction execution system such as, for example, a processor 1303 in a computer system or other system.
  • the logic may comprise, for example, statements including instructions and declarations that can be fetched from the computer-readable medium and executed by the instruction execution system.
  • a “computer-readable medium” can be any medium that can contain, store, or maintain the logic or application described herein for use by or in connection with the instruction execution system.
  • the computer-readable medium can comprise any one of many physical media such as, for example, magnetic, optical, or semiconductor media. More specific examples of a suitable computer-readable medium would include, but are not limited to, magnetic tapes, magnetic floppy diskettes, magnetic hard drives, memory cards, solid-state drives, USB flash drives, or optical discs. Also, the computer-readable medium may be a random access memory (RAM) including, for example, static random access memory (SRAM) and dynamic random access memory (DRAM), or magnetic random access memory (MRAM).
  • RAM random access memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • MRAM magnetic random access memory
  • the computer-readable medium may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or other type of memory device.
  • ROM read-only memory
  • PROM programmable read-only memory
  • EPROM erasable programmable read-only memory
  • EEPROM electrically erasable programmable read-only memory
  • any logic or application described herein, including the device fingerprinting program 1315 and the application program 1318 may be implemented and structured in a variety of ways.
  • one or more applications described may be implemented as modules or components of a single application.
  • one or more applications described herein may be executed in shared or separate computing devices or a combination thereof.
  • a plurality of the applications described herein may execute in the same device 1300 , or in multiple computing devices in the same control system environment 100 .
  • terms such as “application,” “service,” “system,” “engine,” “module,” and so on may be interchangeable and are not intended to be limiting.
  • ratios, concentrations, amounts, and other numerical data may be expressed herein in a range format. It is to be understood that such a range format is used for convenience and brevity, and thus, should be interpreted in a flexible manner to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited.
  • a concentration range of “about 0.1% to about 5%” should be interpreted to include not only the explicitly recited concentration of about 0.1 wt % to about 5 wt %, but also include individual concentrations (e.g., 1%, 2%, 3%, and 4%) and the sub-ranges (e.g., 0.5%, 1.1%, 2.2%, 3.3%, and 4.4%) within the indicated range.
  • the term “about” can include traditional rounding according to significant figures of the numerical value.
  • the phrase “about ‘x’ to ‘y’” includes “about ‘x’ to about ‘y”’.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Environmental & Geological Engineering (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
US15/556,136 2015-03-06 2016-03-04 Device fingerprinting for cyber-physical systems Abandoned US20180048550A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/556,136 US20180048550A1 (en) 2015-03-06 2016-03-04 Device fingerprinting for cyber-physical systems

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US201562129382P 2015-03-06 2015-03-06
US201562202262P 2015-08-07 2015-08-07
US15/556,136 US20180048550A1 (en) 2015-03-06 2016-03-04 Device fingerprinting for cyber-physical systems
PCT/US2016/020985 WO2016144793A1 (fr) 2015-03-06 2016-03-04 Génération d'empreintes digitales de dispositifs pour des systèmes cyber-physiques

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2016/020985 A-371-Of-International WO2016144793A1 (fr) 2015-03-06 2016-03-04 Génération d'empreintes digitales de dispositifs pour des systèmes cyber-physiques

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/583,988 Continuation US11228517B2 (en) 2015-03-06 2019-09-26 Device fingerprinting for cyber-physical systems

Publications (1)

Publication Number Publication Date
US20180048550A1 true US20180048550A1 (en) 2018-02-15

Family

ID=56878992

Family Applications (2)

Application Number Title Priority Date Filing Date
US15/556,136 Abandoned US20180048550A1 (en) 2015-03-06 2016-03-04 Device fingerprinting for cyber-physical systems
US16/583,988 Active 2036-08-02 US11228517B2 (en) 2015-03-06 2019-09-26 Device fingerprinting for cyber-physical systems

Family Applications After (1)

Application Number Title Priority Date Filing Date
US16/583,988 Active 2036-08-02 US11228517B2 (en) 2015-03-06 2019-09-26 Device fingerprinting for cyber-physical systems

Country Status (3)

Country Link
US (2) US20180048550A1 (fr)
EP (1) EP3265919B1 (fr)
WO (1) WO2016144793A1 (fr)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190089741A1 (en) * 2017-09-18 2019-03-21 Veracity Security Intelligence, Inc. Network asset characterization, classification, grouping and control
US20190191009A1 (en) * 2017-12-15 2019-06-20 Sap Se Network based machine learning generated simulations
US10367846B2 (en) * 2017-11-15 2019-07-30 Xm Cyber Ltd. Selectively choosing between actual-attack and simulation/evaluation for validating a vulnerability of a network node during execution of a penetration testing campaign
CN110401662A (zh) * 2019-07-29 2019-11-01 华能阜新风力发电有限责任公司 一种工控设备指纹识别方法、存储介质
US20200153694A1 (en) * 2018-11-13 2020-05-14 Cisco Technology, Inc. Removal of environment and local context from network traffic for device classification
EP3674942A1 (fr) * 2018-12-28 2020-07-01 AO Kaspersky Lab Système et procédé d'identification d'une activité frauduleuse à partir d'un dispositif utilisateur à l'aide d'une chaîne d'empreintes digitales de dispositif
CN111382417A (zh) * 2018-12-28 2020-07-07 卡巴斯基实验室股份制公司 使用一系列设备指纹识别来自用户设备的欺诈行为的系统和方法
US10965675B2 (en) 2018-03-14 2021-03-30 Bank Of America Corporation Preventing unauthorized access to secure information systems using advanced pre-authentication techniques
US11140183B2 (en) * 2019-01-29 2021-10-05 EMC IP Holding Company LLC Determining criticality of identified enterprise assets using network session information
CN117675755A (zh) * 2024-01-31 2024-03-08 浙江省电子信息产品检验研究院(浙江省信息化和工业化融合促进中心) 智能网联设备管理方法与装置

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112073375B (zh) * 2020-08-07 2023-09-26 中国电力科学研究院有限公司 一种适用于电力物联网客户侧的隔离装置及隔离方法
CN112953961B (zh) * 2021-03-14 2022-05-17 国网浙江省电力有限公司电力科学研究院 配电房物联网中设备类型识别方法
DE102022133826A1 (de) 2022-12-19 2024-06-20 Endress+Hauser SE+Co. KG Verfahren und System zum gegenseitigen Überprüfen der Integrität einer Vielzahl von Feldgeräten der Automatisierungstechnik

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020159396A1 (en) * 2001-04-25 2002-10-31 Carlson David G. Adaptive TCP delayed acknowledgment
US20080059636A1 (en) * 2001-06-27 2008-03-06 Freimuth Douglas M In-kernel content-aware service differentiation
US7929508B1 (en) * 2008-06-11 2011-04-19 Atheros Communications, Inc. Radio frequency signal analysis and classification using time-frequency information
US20120019395A1 (en) * 2010-02-22 2012-01-26 Enernoc, Inc. Apparatus and method for network-based grid management
US20120198047A1 (en) * 2011-01-27 2012-08-02 Steuer Rotem Method and system for determining response time of a server
US20130139263A1 (en) * 2011-11-29 2013-05-30 Georgia Tech Research Corporation Systems and methods for fingerprinting physical devices and device types based on network traffic
US20130242795A1 (en) * 2010-11-25 2013-09-19 Thomson Licensing Method and device for fingerprinting of wireless communication devices
US20140173220A1 (en) * 2009-10-30 2014-06-19 Netapp, Inc. Using Logical Block Addresses with Generation Numbers as Data Fingerprints to Provide Cache Coherency
US20150026374A1 (en) * 2013-07-19 2015-01-22 International Business Machines Corporation Managing slave devices
US20150378339A1 (en) * 2014-06-27 2015-12-31 Siemens Aktiengesellschaft Resilient control design for distributed cyber-physical systems
US20170041205A1 (en) * 2015-08-07 2017-02-09 Drayson Technologies (Europe) Limited Power Efficient Control and Operation of a Data-Sensing Peripheral Device Based on Location and Mode of Transport

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4176341B2 (ja) * 2001-10-23 2008-11-05 株式会社日立製作所 記憶制御装置
JP3931823B2 (ja) * 2002-03-18 2007-06-20 セイコーエプソン株式会社 圧電アクチュエータ及びその製造方法、並びに、液体噴射ヘッド及びその製造方法
GB0313002D0 (en) * 2003-06-06 2003-07-09 Ncr Int Inc Currency validation
EP1719114A2 (fr) * 2004-02-18 2006-11-08 Philips Intellectual Property & Standards GmbH Procede et systeme permettant de produire des donnees de formation pour un dispositif de reconnaissance automatique de la parole
US7908020B2 (en) 2004-12-24 2011-03-15 Donald Pieronek Architecture for control systems
US8863293B2 (en) * 2012-05-23 2014-10-14 International Business Machines Corporation Predicting attacks based on probabilistic game-theory

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020159396A1 (en) * 2001-04-25 2002-10-31 Carlson David G. Adaptive TCP delayed acknowledgment
US20080059636A1 (en) * 2001-06-27 2008-03-06 Freimuth Douglas M In-kernel content-aware service differentiation
US7929508B1 (en) * 2008-06-11 2011-04-19 Atheros Communications, Inc. Radio frequency signal analysis and classification using time-frequency information
US20140173220A1 (en) * 2009-10-30 2014-06-19 Netapp, Inc. Using Logical Block Addresses with Generation Numbers as Data Fingerprints to Provide Cache Coherency
US20120019395A1 (en) * 2010-02-22 2012-01-26 Enernoc, Inc. Apparatus and method for network-based grid management
US20130242795A1 (en) * 2010-11-25 2013-09-19 Thomson Licensing Method and device for fingerprinting of wireless communication devices
US20120198047A1 (en) * 2011-01-27 2012-08-02 Steuer Rotem Method and system for determining response time of a server
US20130139263A1 (en) * 2011-11-29 2013-05-30 Georgia Tech Research Corporation Systems and methods for fingerprinting physical devices and device types based on network traffic
US20150026374A1 (en) * 2013-07-19 2015-01-22 International Business Machines Corporation Managing slave devices
US20150378339A1 (en) * 2014-06-27 2015-12-31 Siemens Aktiengesellschaft Resilient control design for distributed cyber-physical systems
US20170041205A1 (en) * 2015-08-07 2017-02-09 Drayson Technologies (Europe) Limited Power Efficient Control and Operation of a Data-Sensing Peripheral Device Based on Location and Mode of Transport

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10742683B2 (en) * 2017-09-18 2020-08-11 Veracity Industrial Networks, Inc. Network asset characterization, classification, grouping and control
US20190089741A1 (en) * 2017-09-18 2019-03-21 Veracity Security Intelligence, Inc. Network asset characterization, classification, grouping and control
US10367846B2 (en) * 2017-11-15 2019-07-30 Xm Cyber Ltd. Selectively choosing between actual-attack and simulation/evaluation for validating a vulnerability of a network node during execution of a penetration testing campaign
US20190191009A1 (en) * 2017-12-15 2019-06-20 Sap Se Network based machine learning generated simulations
US10693997B2 (en) * 2017-12-15 2020-06-23 Sap Se Network based machine learning generated simulations
US10965675B2 (en) 2018-03-14 2021-03-30 Bank Of America Corporation Preventing unauthorized access to secure information systems using advanced pre-authentication techniques
US20200153694A1 (en) * 2018-11-13 2020-05-14 Cisco Technology, Inc. Removal of environment and local context from network traffic for device classification
US10826772B2 (en) * 2018-11-13 2020-11-03 Cisco Technology, Inc. Removal of environment and local context from network traffic for device classification
EP3674942A1 (fr) * 2018-12-28 2020-07-01 AO Kaspersky Lab Système et procédé d'identification d'une activité frauduleuse à partir d'un dispositif utilisateur à l'aide d'une chaîne d'empreintes digitales de dispositif
CN111382417A (zh) * 2018-12-28 2020-07-07 卡巴斯基实验室股份制公司 使用一系列设备指纹识别来自用户设备的欺诈行为的系统和方法
US10931697B2 (en) 2018-12-28 2021-02-23 AO Kaspersky Lab System and method of identifying fraudulent activity from a user device using a chain of device fingerprints
US11140183B2 (en) * 2019-01-29 2021-10-05 EMC IP Holding Company LLC Determining criticality of identified enterprise assets using network session information
CN110401662A (zh) * 2019-07-29 2019-11-01 华能阜新风力发电有限责任公司 一种工控设备指纹识别方法、存储介质
CN117675755A (zh) * 2024-01-31 2024-03-08 浙江省电子信息产品检验研究院(浙江省信息化和工业化融合促进中心) 智能网联设备管理方法与装置

Also Published As

Publication number Publication date
EP3265919A4 (fr) 2019-03-20
EP3265919B1 (fr) 2021-09-29
WO2016144793A1 (fr) 2016-09-15
US11228517B2 (en) 2022-01-18
EP3265919A1 (fr) 2018-01-10
US20200106686A1 (en) 2020-04-02

Similar Documents

Publication Publication Date Title
US11228517B2 (en) Device fingerprinting for cyber-physical systems
Formby et al. Who's in Control of Your Control System? Device Fingerprinting for Cyber-Physical Systems.
Wang et al. Anomaly detection for industrial control system based on autoencoder neural network
Singh et al. A cyber-physical anomaly detection for wide-area protection using machine learning
Li et al. System statistics learning-based IoT security: Feasibility and suitability
Bertoli et al. An end-to-end framework for machine learning-based network intrusion detection system
Lemay et al. Providing {SCADA} network data sets for intrusion detection research
Radoglou-Grammatikis et al. Spear siem: A security information and event management system for the smart grid
US11689544B2 (en) Intrusion detection via semantic fuzzing and message provenance
Yang et al. Robust detection of false data injection attacks for data aggregation in an Internet of Things-based environmental surveillance
US9367683B2 (en) Cyber security
Regan et al. Federated IoT attack detection using decentralized edge data
Kukkala et al. Latte: L stm self-att ention based anomaly detection in e mbedded automotive platforms
Marsden et al. Probability risk identification based intrusion detection system for SCADA systems
Zheng et al. Smart grid: Cyber attacks, critical defense approaches, and digital twin
Cepeda et al. Data‐mining‐based approach for predicting the power system post‐contingency dynamic vulnerability status
Ahmed et al. Scanning the cycle: Timing-based authentication on PLCs
Alrimawi et al. Software engineering challenges for investigating cyber-physical incidents
Adiban et al. A step-by-step training method for multi generator GANs with application to anomaly detection and cybersecurity
Sarieddine et al. Edge-based detection and localization of adversarial oscillatory load attacks orchestrated by compromised EV charging stations
Wang et al. A novel energy-based online sequential extreme learning machine to detect anomalies over real-time data streams
Kadri et al. Survey and classification of Dos and DDos attack detection and validation approaches for IoT environments
Babun et al. The Truth Shall Set Thee Free: Enabling Practical Forensic Capabilities in Smart Environments.
Onoda Probabilistic models-based intrusion detection using sequence characteristics in control system communication
Efiong et al. A contrived dataset of substation automation for cybersecurity research in the smart grid networks based on IEC61850

Legal Events

Date Code Title Description
AS Assignment

Owner name: GEORGIA TECH RESEARCH CORPORATION, GEORGIA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:BEYAH, ABDUL RAHEEM;FORMBY, DAVID;SRINIVASAN, PREETHI;REEL/FRAME:044281/0699

Effective date: 20170830

AS Assignment

Owner name: NATIONAL SCIENCE FOUNDATION, VIRGINIA

Free format text: CONFIRMATORY LICENSE;ASSIGNOR:GEORGIA INSTITUTE OF TECHNOLOGY;REEL/FRAME:048356/0169

Effective date: 20190204

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: FINAL REJECTION MAILED

STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION