Content distributed network, CDN, controller and a method for minimizing impact of one or more rate limiters
The CDN controller addresses the challenge of balancing security and usability by using traffic histograms and Attack Differentiating Factors to dynamically adjust rate limiters, enhancing user experience and resilience against attacks while minimizing impact on legitimate traffic.
Patent Information
- Application Number
- PCT/EP2023/062869
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-15
- Publication Date
- 2024-11-21
AI Technical Summary
Traditional traffic enforcement techniques like rate limiting and traffic shaping often negatively impact user experience by causing interruptions and slower page load times, making it challenging to balance security and usability in network protection against threats such as DDoS attacks.
A Content Distributed Network (CDN) controller that generates traffic histograms and Attack Differentiating Factors to accurately differentiate between benign and malicious traffic, allowing for dynamic adjustment of rate limiters to minimize impact on legitimate users while effectively mitigating attacks.
The CDN controller provides more accurate filtering, reduces false positives, and enhances user experience by ensuring effective attack mitigation without impacting legitimate traffic, improving resilience against zero-day and sophisticated attacks with real-time efficiency and simplicity.
Smart Images

Figure EP2023062869_21112024_PF_FP_ABST
Abstract
Description
[0001] CONTENT DISTRIBUTED NETWORK, CDN, CONTROLLER AND A METHOD FOR MINIMIZING IMPACT OF ONE OR MORE RATE LIMITERS
[0002] TECHNICAL FIELD
[0003] The disclosure relates generally to a cyber-security, and more particularly, the disclosure relates to a Content Distributed Network, CDN, controller, and a method for minimizing impact of one or more rate limiters.
[0004] BACKGROUND
[0005] The challenge of balancing security and usability in the context of traffic enforcement arises from the fact that many traditional traffic enforcement techniques, such as rate limiting, traffic shaping, or packet filtering, can have negative effects on user experience.
[0006] Packet drop filters are designed to block certain packets from reaching their destination, but they also cause interruptions in legitimate sessions and retransmissions, leading to a poor user experience. To minimize the impact, it's important to carefully tune the filters and set them to appropriate thresholds. Network administrators have to monitor the impact on user experience and make adjustments as needed.
[0007] These techniques aim to restrict or prioritise network traffic based on certain criteria, such as source Internet Protocol, IP, address, packet type, or time of day, in order to prevent various threats such as Distributed Denial of Service, DDoS, attacks, malware infections, or unauthorised access.
[0008] These techniques can also result in slower page load times, increased latency, or other types of interference that can degrade the user experience. This can lead to frustrated users, lost revenue, or reduced productivity, which can ultimately undermine the overall effectiveness of the network or system. The challenge of balancing security and usability is especially acute in today's digital landscape, where users expect fast, reliable, and seamless access to information and services. In order to remain competitive, organisations must find ways to protect their networks and systems against threats while also providing a positive user experience. A more nuanced, context-aware approach to traffic enforcement that considers user behaviour, device type, and other factors. For example, rather than simply rate limiting all traffic from a particular source Internet Protocol, IP, address, administrators might use machine learning techniques to identify patterns of behaviour that indicate malicious intent, and only restrict traffic that falls into these patterns. The need for this method arises from the fact that many traditional traffic enforcement techniques, such as rate limiting or traffic shaping, can negatively impact the user experience by slowing down page load times or otherwise interfering with legitimate traffic. By estimating the impact on user experience, network administrators can make more informed decisions about which enforcement techniques to use and how to configure them.
[0009] Combining multiple traffic enforcement mechanisms in a way that minimises negative impact on user experience while still providing effective protection against various threats. The need for this principle arises from the fact that no single traffic enforcement technique is perfect, and different techniques may be better suited to different types of threats. By combining multiple techniques in an optimised way, network administrators can provide better protection against threats without sacrificing user experience. The traffic enforcement ensemble is adjusted based on feedback from the network or system. The need for this method arises from the fact that traffic patterns and threats can change over time, and a static configuration may not be optimal in all circumstances. By automatically adjusting the ensemble structure based on feedback, network administrators can ensure that the system remains both secure and usable.
[0010] Therefore, there arises a need to address the aforementioned technical problem / drawbacks in balancing security and usability in the context of traffic enforcement.
[0011] SUMMARY
[0012] It is an object of the disclosure to provide a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters and a method for use in the CDN controller for minimizing impact of one or more rate limiters while avoiding one or more disadvantages of prior art approaches.
[0013] This object is achieved by the features of the independent claims. Further, implementation forms are apparent from the dependent claims, the description, and the figures. The disclosure provides a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters and a method for use in the CDN controller for minimizing impact of one or more rate limiters.
[0014] According to a first aspect, there is provided a Content Distributed Network, CDN, controller. The CDN controller is configured to minimize impact of one or more rate limiters. The CDN controller is configured to receive a first plurality of data packets. Each data packet including one or more packet features. The CDN controller is configured to generate a traffic histogram for the packet feature based on the first plurality of data packets. The histogram for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the traffic histogram. The traffic histogram indicates benign traffic quotas, BQ. The CDN controller is configured to determine that there is an attack undergoing. The CDN controller is configured to receive a second plurality of data packets. The CDN controller is configured to determine an effect of a rate limiter on a packet feature by (a) generating an attack histogram for the packet feature based on the plurality of data packets and (b) determining an Attack Differentiating Factor, ADF, for each bin, in the attack histogram based on the formula ADF (f, i) = AQ(f, i) / BQ(f, i), where f indicates the packet feature and i indicates the histogram value bin. The attack histogram indicates attack traffic quotas, AQ. The Attack Differentiating Factor, ADF, indicates the effect of the rate limiter.
[0015] The CDN controller provides more accurate filtering techniques. The CDN controller minimises the impact on legitimate users, and reduces false positives. The CDN controller improves the user experience by ensuring that attacks are effectively mitigated without impacting the availability of legitimate traffic. A system that is implemented with the CDN controller improves resiliency against zero day and sophisticated attacks with relative simplicity, and no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency. The CDN controller is highly detectable, allowing for easy evaluation of the system's handling of adversarial traffic. The CDN controller may also be applied on a cloud infrastructure. The CDN controller offers a more effective and user-friendly way to mitigate Distributed Denial of Service, DDoS, attacks.
[0016] The CDN controller improves a reliability of content delivery. By distributing content across multiple servers and geographic locations, the CDN controller ensures that the content remains available even in the event of a server failure or network outage. The CDN controller easily scale to handle increasing amounts of traffic, allowing organisations to accommodate spikes in demand without experiencing performance issues. The CDN controller reduces the load on origin servers by handling requests for cacheable content, which improves server performance and reduce the risk of server overload during peak traffic periods. The CDN controller provides better security for content delivery by offering Distributed Denial of Service, DDoS, protection, Secure Socket Layers, SSL, encryption, and other security features.
[0017] Optionally, the CDN controller is further configured to determine the Attack Differentiating Factor, ADF, based on the formula ADF (f, i) = AQ(f, i) / Max(BQ(f, i); MinBQ(f, i)), where MinBQ(f, i) is a set minimum value for the bin i for feature f during traffic prior to the attack.
[0018] The Attack Differentiating Factor, ADF, distinguishes legitimate traffic from malicious traffic during an attack. By minimizing the Attack Differentiating Factor, ADF, it is possible to identify when an attack is occurring and adjust the rate limiter accordingly.
[0019] Optionally, the CDN controller is further configured to set the MinBQ to a default minimum value.
[0020] Optionally, the CDN controller is further configured to select a rate limiter to be combined from a plurality of bin specific rate limiters. Each rate limiter is associated with one or more bin of appropriate feature’s histogram. The CDN controller is configured to determine the Attack Differentiating Factor, ADF, for one or more packet features. The CDN controller is configured to select the rate limiter to be enforced to be a rate limiter associated with the bin of packet feature having the highest Attack Differentiating Factor, ADF.
[0021] Optionally, the CDN controller is further configured to determine that there is a packet feature having an Attack Differentiating Factor, ADF, exceeding a threshold value prior to selecting the rate limiter to be enforced.
[0022] Optionally, the controller is further configured to determine that an ADF exceeds a threshold value utilizing a Kolmogorov- Smirnov Test. Optionally, the CDN controller is further configured to determine that there is no packet feature having an Attack Differentiating Factor, ADF, exceeding a threshold value prior to selecting the rate limiter to be enforced, and in response thereto not selecting a rate limiter to be enforced.
[0023] The rate limiters may reduce the impact on legitimate traffic during an attack, ensuring that normal network operations can still take place. The rate limiters provide a quick response time, enabling organisations to detect and mitigate attacks in real-time. The rate limiters provide several advantages for organisations facing cyber-attacks. The rate limiters protect against volumetric attacks and prevent overwhelming traffic volumes that may cause a denial of service.
[0024] Optionally, the CDN controller is further configured to determine if a further rate limiter to be enforced is to be selected by (a) determining a traffic volume to be removed, RMV, and (b) determining that the traffic volume to be removed, RMV, is above 0 and in response thereto select a further rate limiter based on the Attack Differentiating Factors, ADFs, determined for the packet features. The traffic volume to be removed is determined as the traffic volume exceeding an attack threshold value.
[0025] Optionally, the CDN controller is further configured to determine that the attack is still ongoing and in response thereto. The CDN controller is configured to receive a further second plurality of data packets and determine a second attack histogram based on the further second plurality of data packets and Attack Differentiating Factors, ADFs, for the packet features.
[0026] Optionally, the CDN controller is further configured to determine that the attack is still ongoing at intervals. The intervals may be at 45, 60, 75, 90, 105, 120, 135, 150, 165 or 180 seconds or any range there in between.
[0027] Optionally, the CDN controller is further configured to determine the traffic volume to be removed, RMV, as RMV = FF x SUM(AQ(f,i)), where FF is a Feedback Factor, determined as FFj = 1- RMVj-1 / RMVj-2, wherein FF0 = FF1= 1.
[0028] Optionally, the packet feature is a nbr packets, a nbr of open connections, a packet size, a header size, payload size, a protocol type, destination port, a source port, a fragment number, a Time to Live, TTL, a Transmission Control Protocol, TCP, flag, a sequence counter, or an Internet Protocol, IP, identifier.
[0029] According to a second aspect, there is provided a method for use in a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters. The method includes receiving a first plurality of data packets, each data packet comprising one or more packet features. The method includes generating a traffic histogram for the packet feature based on the first plurality of data packets. The histogram for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the traffic histogram. The traffic histogram indicates benign traffic quotas, BQ. The method includes determining that there is an attack undergoing. The method includes receiving a second plurality of data packets. The method includes determining an effect of a rate limiter on a packet feature by generating an attack histogram for the packet feature based on the plurality of data packets. The attack histogram indicates attack traffic quotas, AQ. The method includes determining an Attack Differentiating Factor, ADF, for each bin, in the attack histogram based on the formula ADF (f, i) = AQ(f, i) / BQ(f, i), where f indicates the packet feature and i indicates the histogram value bin. The Attack Differentiating Factor, ADF, indicates the effect of the rate limiter.
[0030] This method provides more accurate filtering techniques. This method minimises the impact on legitimate users, and reduces false positives. This method improves the user experience by ensuring that attacks are effectively mitigated without impacting the availability of legitimate traffic. This method improves system resiliency against zero day and sophisticated attacks with relative simplicity, and no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency. This method is highly detectable, allowing for easy evaluation of the system's handling of adversarial traffic. This method offers a more effective and user-friendly way to mitigate Distributed Denial of Service, DDoS, attacks.
[0031] Optionally, the method includes selecting a rate limiter to be combined from a plurality of bin specific rate limiters. Each rate limiter is associated with one or more bin of appropriate feature’s histogram. The method includes determining the Attack Differentiating Factor, ADF, for one or more packet features. The method includes selecting the rate limiter to be enforced to be a rate limiter associated with the bin of packet feature having the highest Attack Differentiating Factor, ADF. According to a third aspect, there is provided a computer program product including program instructions for performing the second aspect, when executed by one or more processors in a Content Distributed Network, CDN, controller system.
[0032] The computer program product provides more accurate filtering techniques. The computer program product minimises the impact on legitimate users, and reduces false positives. The computer program product improves the user experience by ensuring that attacks are effectively mitigated without impacting the availability of legitimate traffic. The computer program product improves system resiliency against zero day and sophisticated attacks with relative simplicity, no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency. The computer program product is highly detectable, allowing for easy evaluation of the system's handling of adversarial traffic. The computer program product offers a more effective and user-friendly way to mitigate Distributed Denial of Service, DDoS, attacks.
[0033] A technical problem in the prior art is resolved, where the technical problem is minimising UX impact, minimizing UX impact and method of feedback-based real-time ensemble readjustments.
[0034] Therefore, in contradistinction to the prior art, according to a CDN controller for minimizing impact of one or more rate limiters and a method of minimizing impact of one or more rate limiters enable more accurate filtering, to improve the overall user experience while minimising the occurrence of True Negative events that are often unavoidable in rate limiters. The CDN controller and the method enhance the system's resiliency against zero-day and sophisticated attacks, making it more secure and robust. The CDN controller and the method provide relative simplicity, no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency, making it an attractive option for organisations seeking to protect against Distributed Denial of Service, DDoS, attacks while maintaining optimal user experience and system performance. The enhancement of Distributed Denial of Service, DDoS, mitigation offers several benefits as it significantly enhances the user experience by implementing more accurate filtering, which minimises the impact of malicious traffic on Quality of Service, QoS, Key Performance Indicators, KPIs such as response times and service disruptions. The number of True Negative events may be reduced, which are unavoidable in rate limiters, by using a more sophisticated system that differentiates more accurately between legitimate and malicious traffic. CDN controller may be made more resilient to zero-day and sophisticated attacks by adapting quickly to detect and mitigate new types of attacks. The CDN controller’s relative simplicity, no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency provides a significant advantage over state-of-the-art tools, providing a cost-effective and easy-to-use solution for organisations seeking to protect their systems and users against cyberattacks. By achieving these advantages, Distributed Denial of Service, DDoS, mitigation may provide businesses and organisations with a reliable and efficient way to safeguard their infrastructure and users against cyber threats.
[0035] These and other aspects of the disclosure will be apparent from and the implementation(s) described below.
[0036] BRIEF DESCRIPTION OF DRAWINGS
[0037] Implementations of the disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which:
[0038] FIG. 1 is a block diagram of a Content Distributed Network, CDN, controller in accordance with an implementation of the disclosure;
[0039] FIGS. 2A to 2D are exemplary graphical views of packet features in a histogram format in accordance with an implementation of the disclosure;
[0040] FIGS. 3A to 3D are exemplary graphical views of packet features in a histogram format for minimising rate limiting rules generation in accordance with an implementation of the disclosure;
[0041] FIG. 3E is an exemplary table view of Attack Differentiating Factor, ADF values of packet features of FIGS. 3A to 3D in accordance with an implementation of the disclosure;
[0042] FIG. 4 is a flow diagram that illustrates a flow of minimizing impact of one or more rate limiters in accordance with an implementation of the disclosure;
[0043] FIGS. 5A and 5B are flow diagrams that illustrate a method of for use in a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters in accordance with an implementation of the disclosure; and
[0044] FIG. 6 is an illustration of a computer system (e.g. a Content Distributed Network, CDN, controller) in which the various architectures and functionalities of the various previous implementations may be implemented. DETAILED DESCRIPTION OF THE DRAWINGS
[0045] Implementations of the disclosure provide a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters and a method for use in a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters.
[0046] To make solutions of the disclosure more comprehensible for a person skilled in the art, the following implementations of the disclosure are described with reference to the accompanying drawings.
[0047] Terms such as "a first", "a second", "a third", and "a fourth" (if any) in the summary, claims, and foregoing accompanying drawings of the disclosure are used to distinguish between similar objects and are not necessarily used to describe a specific sequence or order. It should be understood that the terms so used are interchangeable under appropriate circumstances, so that the implementations of the disclosure described herein are, for example, capable of being implemented in sequences other than the sequences illustrated or described herein. Furthermore, the terms "include" and "have" and any variations thereof, are intended to cover a non-ex elusive inclusion. For example, a process, a method, a system, a product, or a device that includes a series of steps or units, is not necessarily limited to expressly listed steps or units but may include other steps or units that are not expressly listed or that are inherent to such process, method, product, or device.
[0048] FIG. 1 is a block diagram of a Content Distributed Network, CDN, controller 102 in accordance with an implementation of the disclosure. The CDN controller 102 is configured to minimize impact of one or more rate limiters. The CDN controller 102 is configured to receive a first plurality of data packets 104A-N. Each data packet includes one or more packet features. The CDN controller 102 is configured to generate a traffic histogram 108 for the packet feature based on the first plurality of data packets 104A-N. The traffic histogram 108 for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the traffic histogram 108. The traffic histogram 108 indicates benign traffic quotas, BQ. The CDN controller 102 is configured to determine that there is an attack undergoing. The CDN controller 102 is configured to receive a second plurality of data packets 106A-N. The CDN controller 102 is configured to determine an effect of a rate limiter on a packet feature by (a) generating an attack histogram 110 for the packet feature based on the plurality of data packets and (b) determining an Attack Differentiating Factor, ADF, for each bin, in the attack histogram 110 based on the formula ADF (f, i) = AQ(f, i) / BQ(f, i), where f indicates the packet feature and i indicates the histogram value bin. The attack histogram 110 indicates attack traffic quotas, AQ. The Attack Differentiating Factor, ADF, indicates the effect of the rate limiter.
[0049] The CDN controller 102 provides more accurate filtering techniques. The CDN controller 102 minimises the impact on legitimate users, and reduces false positives. The CDN controller 102 improves the user experience by ensuring that attacks are effectively mitigated without impacting the availability of legitimate traffic. A system that is implemented with the CDN controller 102 improves resiliency against zero day and sophisticated attacks with relative simplicity, and no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency. The CDN controller 102 is highly detectable, allowing for easy evaluation of the system's handling of adversarial traffic. The CDN controller 102 may also be applied on a cloud infrastructure. The CDN controller 102 offers a more effective and user-friendly way to mitigate Distributed Denial of Service, DDoS, attacks.
[0050] The CDN controller 102 improves a reliability of content delivery. By distributing content across multiple servers and geographic locations, the CDN controller 102 ensures that the content remains available even in the event of a server failure or network outage. The CDN controller 102 easily scale to handle increasing amounts of traffic, allowing organisations to accommodate spikes in demand without experiencing performance issues. The CDN controller 102 reduces the load on origin servers by handling requests for cacheable content, which improves server performance and reduce the risk of server overload during peak traffic periods. The CDN controller 102 provides better security for content delivery by offering Distributed Denial of Service, DDoS, protection, Secure Socket Layers, SSL, encryption, and other security features.
[0051] Optionally, the CDN controller 102 is configured to determine the Attack Differentiating Factor, ADF, based on the formula ADF (f, i) = AQ(f, i) / Max(BQ(f, i); MinBQ(f, i)), where MinBQ(f, i) is a set minimum value for the bin i for feature f during traffic prior to the attack. The Attack Differentiating Factor, ADF, distinguishes legitimate traffic from malicious traffic during an attack. By minimizing the Attack Differentiating Factor, ADF, it is possible to identify when an attack is occurring and adjust the rate limiter accordingly.
[0052] Optionally, the CDN controller 102 is configured to set the MinBQ to a default minimum value.
[0053] Optionally, the CDN controller 102 is configured to select a rate limiter to be combined from a plurality of bin specific rate limiters. Each rate limiter is associated with one or more bin of appropriate feature’s histogram. The CDN controller 102 is configured to determine the Attack Differentiating Factor, ADF, for one or more packet features. The CDN controller 102 is configured to select the rate limiter to be enforced to be a rate limiter associated with the bin of packet feature having the highest Attack Differentiating Factor, ADF.
[0054] Optionally, the CDN controller 102 is further configured to determine that there is a packet feature having an Attack Differentiating Factor, ADF, exceeding a threshold value prior to selecting the rate limiter to be enforced. Optionally, the CDN controller 102 is further configured to determine that an ADF exceeds a threshold value utilizing a Kolmogorov- Smirnov Test. Optionally, the Kolmogorov- Smirnov Test defines the maximum possible difference between threshold and current value that still considered insignificant (not breaking the normal parameter deviations).
[0055] Optionally, the CDN controller 102 is further configured to determine that there is no packet feature having an Attack Differentiating Factor, ADF, exceeding a threshold value prior to selecting the rate limiter to be enforced, and in response thereto not selecting a rate limiter to be enforced.
[0056] The rate limiters may reduce the impact on legitimate traffic during an attack, ensuring that normal network operations can still take place. The rate limiters provide a quick response time, enabling organisations to detect and mitigate attacks in real-time. The rate limiters provide several advantages for organisations facing cyber-attacks. The rate limiters protect against volumetric attacks and prevent overwhelming traffic volumes that may cause a denial of service. Optionally, the CDN controller 102 is further configured to determine if a further rate limiter to be enforced is to be selected by (a) determining a traffic volume to be removed, RMV, and (b) determining that the traffic volume to be removed, RMV, is above 0 and in response thereto select a further rate limiter based on the Attack Differentiating Factors, ADFs, determined for the packet features. The traffic volume to be removed is determined as the traffic volume exceeding an attack threshold value.
[0057] Optionally, the CDN controller 102 is further configured to determine that the attack is still ongoing and in response thereto. The CDN controller 102 is configured to receive a further second plurality of data packets and determine a second attack histogram based on the further second plurality of data packets and Attack Differentiating Factors, ADFs, for the packet features.
[0058] Optionally, the CDN controller 102 is further configured to determine that the attack is still ongoing at intervals. The intervals may be at 45, 60, 75, 90, 105, 120, 135, 150, 165 or 180 seconds or any range there in between.
[0059] Optionally, the CDN controller 102 is further configured to determine the traffic volume to be removed, RMV, as RMV = FF x SUM(AQ(f,i)), where FF is a Feedback Factor, determined as FFj = 1- RMVj-1 / RMVj-2, wherein FF0 = FF1= 1.
[0060] Optionally, the packet feature is a nbr packets, a nbr of open connections, a packet size, a header size, payload size, a protocol type, destination port, a source port, a fragment number, a Time to Live, TTL, a Transmission Control Protocol, TCP, flag, a sequence counter, or an Internet Protocol, IP, identifier.
[0061] FIGS. 2A to 2D are exemplary graphical views of packet features in a histogram format in accordance with an implementation of the disclosure. The packet feature are a nbr packets, a nbr of open connections, a packet size, a header size, payload size, a protocol type, destination port, a source port, a fragment number, a Time to Live, TTL, a Transmission Control Protocol, TCP, flag, a sequence counter, or an Internet Protocol, IP, identifier. The packet features are represented as histograms against values in X axis with corresponding percentage in Y axis. Optionally, the packet features are provided to create histograms utilised in anomaly detection and generating optimal rate limiting rules. The packet features include one or more graphical views for at least one of a packet size 202A, Transmission Control Protocol, TCP flags 202B, a Time to Live, TTL, 202C, and source ports 202D. Optionally, the packet features include one or more graphical representations for an Internet Protocol, IP, header size, a payload size, a L4 header size, a L4 payload size, a protocol type, a destination port, a source port, a geo-location, a fragment number, a sequence number or an Internet Protocol, IP, Identity, ID. The histograms may show signs of a distribution breakage under Distributed Denial of Service, DDoS, attacks. The signs may include high values of the bins compared with expected values. The histogram with broken distribution may be selected for evaluation of user experience, UX impact and rate limiting rules generation.
[0062] FIGS. 3A to 3D are exemplary graphical views of packet features in a histogram format for minimizing rate limiting rules generation in accordance with an implementation of the disclosure. A graphical view of a histogram 302 includes adversarial traffic and benign traffic. The histogram 302 includes one or more packet features and one or more bins i.e. VI to V5. An Attack Differentiating Factor, ADF, is calculated for the histogram 302 for all the bins using a formula ADF (f, i) = AQ(f, i) / Max(BQ(f, i); MinBQ(f, i)), where MinBQ(f, i) is a set minimum value for the bin i for feature f during traffic prior to the attack. Optionally, the f represents a feature identifier, the i represents a number of the bin in f related histogram 302, the AQ represents a quote of the adversarial traffic in the overall incoming traffic, the BQ represents a quote of the benign traffic in the overall incoming traffic. The MinBQ may be a minimal possible value of the appropriate bin which is greater than 0. This is measured in integers in the range 1 - 1..10000 of packets or bytes.
[0063] The histogram 302 includes a bin 304 i.e. the V4, with highest Attack Differentiating Factor, ADF, that is limiting rate using specific F = V4 condition may provide highest suppression of the adversarial traffic almost without creating an impact on the benign traffic, which may not affect user experience, UX. The bins of the selected histograms may be sorted by Attack Differentiating Factor, ADF.
[0064] Top ranked bins may be used for creating Rate Limiting Rules, RLM, rules 308 for traffic volume to be removed, RMV, 306 to remove excessive traffic and bringing filtered traffic back to the normal area under the threshold, as shown in graphical views of the FIGS. 3B, 3C and 3D. The Rate Limiting Rules, RLM, rules 308 may overlap, but overall, the traffic volume to be removed, RMV, 306 created by all the Rate Limiting Rules, RLM, rules 308 may be smaller than the absolute sum. The method of the feedback based self-optimising rule ensembling Feedback Factor, FF, which is calculated as FF = 1 - Recent RMV / Previous RMV to overcome the Rate Limiting Rules, RLM, rules 308 overlap issue.
[0065] FIG. 3E is an exemplary table view 310 of Attack Differentiating Factor, ADF values of packet features of FIGS. 3 A to 3D in accordance with an implementation of the disclosure. The table view 310 includes number of buckets i.e. bins including VI to V5, and corresponding values associated with benign traffic quotas, BQ, attack traffic quotas, AQ and Attack Differentiating Factor, ADF.
[0066] Dropping of valid packets by rate limiting filters may have an effect on the user experience, UX. An attack vector, i.e., a Transmission control Protocol, TCP, packets flood, is recognised as excessive traffic produced by the appropriate type packets and generates a rule to drop packets. The attack vector may eliminate the necessary excessive traffic. For instance, if 70% of the Transmission control Protocol, TCP, traffic is dropped, 7 out of 10 packets with Transmission control Protocol, TCP, values in the Protocol Type fields will be dropped, which results in the majority of discarded packets as adversarial and statistically dropped packets belongs to benign traffic. Transmission control Protocol, TCP, sessions with dropped packets provide sluggishness with additional retransmission, or terminates when several packets from the same session fail to exchange. A probability of dropping benign packets is lower in the selection than in alternatives as the benign traffic has lowest percentage, which minimizes UX impact and provides a way to selectively drop packets. The rate between attack quotes and benign quotes in traffic with the proper feature including TCP Packets, Packets of size 76, Packets with 1213 destination port, and the like, is known as the Attack Differentiating Factor, or the ADF. The Attack Differentiating Factor, ADF, and crate drop rules to rank the appropriate packet features based on highly rated bins. In the highest-ranking bin, dropping packets from bin 4 i.e. V4 statistically results in just 2% of the valid packets being lost. According to statistics, a bin 2 i.e. V2 dropped packets result in 20% of the valid packets being lost, which negatively affects UX. Optionally, when a flooding portion of traffic cannot be cleaned using a single filtering rule with a rank of bin 4 i.e. V4 may be the bin 2 i.e. V2 can be used.
[0067] FIG. 4 is a flow diagram that illustrates a flow of minimizing impact of one or more rate limiters in accordance with an implementation of the disclosure. A first set of first plurality of data packets are received in a system implemented with a Content Distributed Network, CDN, controller. Optionally, each data packet includes one or more packet features. Optionally, based on the first set of plurality of data packets a traffic histogram for the packet feature is created. At a step 402, an undergoing attack is determined. The undergoing attack may be a Distributed Denial of Service, DDoS, attack. Optionally, the Distributed Denial of Service, DDoS, attack is detected through a significant decrease in one or more Quality of Service, QoS, Key Performance Indicators, KPIs, including response times or service disruptions. Optionally, technical measurements, including Traffic Packets Per Second, PPS, and Bits Per Second, BPS, Central Processing Unit, CPU, percentage, and Number of Open Transmission Control Protocol, TCP, Connections, identifies the DDoS attack. At a step 404, distributed denial of service, DDoS indicators are evaluated when the attack is determined. At a step 406, based on the distributed denial of service, DDoS indicators evaluations, whether the Key Performance Indicators, KPI, are in limit or not is determined. If the KPI is in limit, the histogram is refreshed, at a step 408. The histogram for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the histogram. The traffic histogram indicates benign traffic quotas, BQ. All the traffic histograms may run concurrently and use pre-recorded traffic snapshots of various depth. At a step 410, reference templates are updated and the steps 406, 408 and 410 are repeated until the KPI limits are exceeded.
[0068] At a step 412, an Attack Differentiating Factor, ADF, is calculated for each bin, in the histogram using a formula, ADF (f, i) = AQ (f, i) / BQ (f, i). The f indicates the packet feature and i indicates the histogram value bin. Optionally, the Attack Differentiating Factor, ADF, indicates an effect of the rate limiter. At a step 414, the bins of the histogram are sorted based on the Attack Differentiating Factor, ADF, in a user experience, UX, rate. At a step 416, traffic volume to be removed, RMV is determined to identify whether the traffic volume to be removed, RMV is reached 0 or not. When the traffic volume to be removed, RMV, is above 0 and in response thereto add Rate Limiting Rules, RLM, on top ranked bins of the histogram, at a step 418. Optionally, a rate limiter based on the ADF is determined for the packet features until the traffic volume to be removed, RMV, is 0. The traffic volume to be removed, RMV is determined as the traffic volume exceeding an attack threshold value. The rate limiter may be based on histograms bins with most effective and least UX intrusive rank. Optionally, if applying “top rank” rate limiter is not bringing QoS KPIs back to normal, the system may add more Rate Limiting Rules, RLM, till reaching required traffic volume to be removed, RMV, goals. When the traffic volume to be removed, RMV, reaches 0, the undergoing attack is determined again, at a step 420. The undergoing attack may be determined when the KPIs are not in limit. During the DDoS attack, the histograms may evaluate continuously to switch to the most efficient in case when attack changes or new histogram becomes available. At a step 422, all the applied rate limiters are removed.
[0069] FIGS. 5A and 5B are flow diagrams that illustrate a method for use in a Content Distributed Network, CDN, controller for minimizing impact of one or more rate limiters in accordance with an implementation of the disclosure. At a step 502, a first plurality of data packets is received. Each data packet includes one or more packet features. At a step 504, a traffic histogram is generated for the packet feature based on the first plurality of data packets. The histogram for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the traffic histogram. The traffic histogram indicates benign traffic quotas, BQ. At a step 506, an undergoing attack is determined. At a step 508, a second plurality of data packets are received. At a step 510, an effect of a rate limiter on a packet feature is determined by (a) generating an attack histogram for the packet feature based on the plurality of data packets, and (b) determining an Attack Differentiating Factor, Attack Differentiating Factor, ADF, for each bin, in the attack histogram based on the formula ADF (f, i) = AQ(f, i) / BQ(f, i), where f indicates the packet feature and i indicates the histogram value bin. The attack histogram indicates attack traffic quotas, AQ. The Attack Differentiating Factor, ADF, indicates the effect of the rate limiter.
[0070] This method provides more accurate filtering techniques. This method minimises the impact on legitimate users, and reduces false positives. This method improves the user experience by ensuring that attacks are effectively mitigated without impacting the availability of legitimate traffic. This method improves system resiliency against zero day and sophisticated attacks with relative simplicity, no need for exhaustive training, applicable to dynamically changing traffic patterns, relatively compact decision engines and real-time efficiency. This method is highly detectable, allowing for easy evaluation of the system's handling of adversarial traffic. This method offers a more effective and user-friendly way to mitigate DDoS attacks.
[0071] Optionally, the method includes selecting a rate limiter to be combined from a plurality of bin specific rate limiters. Each rate limiter is associated with one or more bin of appropriate feature’s histogram. The method includes determining the Attack Differentiating Factor, ADF, for one or more packet features. The method includes selecting the rate limiter to be enforced to be a rate limiter associated with the bin of packet feature having the highest Attack Differentiating Factor, ADF.
[0072] In an implementation, there is provided a computer program product including program instructions for performing the method, when executed by one or more processors in a Content Distributed Network, CDN, controller system.
[0073] FIG. 6 is an illustration of a computer system (e.g. a database management system) in which the various architectures and functionalities of the various previous implementations may be implemented. As shown, the computer system 600 includes at least one processor 604 that is connected to a bus 602, wherein the computer system 600 may be implemented using any suitable protocol, such as PCI (Peripheral Component Interconnect), PCI-Express, AGP (Accelerated Graphics Port), Hyper Transport, or any other bus or point-to-point communication protocol (s). The computer system 600 also includes a memory 606.
[0074] Control logic (software) and data are stored in the memory 606 which may take a form of random-access memory (RAM). In the disclosure, a single semiconductor platform may refer to a sole unitary semiconductor-based integrated circuit or chip. It should be noted that the term single semiconductor platform may also refer to multi-chip modules with increased connectivity which simulate on-chip modules with increased connectivity which simulate on- chip operation, and make substantial improvements over utilising a conventional central processing unit (CPU) and bus implementation. Of course, the various modules may also be situated separately or in various combinations of semiconductor platforms per the desires of the user.
[0075] The computer system 600 may also include a secondary storage 610. The secondary storage 610 includes, for example, a hard disk drive and a removable storage drive, representing a floppy disk drive, a magnetic tape drive, a compact disk drive, digital versatile disk (DVD) drive, recording device, universal serial bus (USB) flash memory. The removable storage drive at least one of reads from and writes to a removable storage unit in a well-known manner.
[0076] Computer programs, or computer control logic algorithms, may be stored in at least one of the memory 606 and the secondary storage 610. Such computer programs, when executed, enable the computer system 600 to perform various functions as described in the foregoing. The memory 606, the secondary storage 610, and any other storage are possible examples of computer-readable media.
[0077] In an implementation, the architectures and functionalities depicted in the various previous figures may be implemented in the context of the processor 604, a graphics processor coupled to a communication interface 612, an integrated circuit (not shown) that is capable of at least a portion of the capabilities of both the processor 604 and a graphics processor, a chipset (namely, a group of integrated circuits designed to work and sold as a unit for performing related functions, and so forth).
[0078] Furthermore, the architectures and functionalities depicted in the various previous-described figures may be implemented in a context of a general computer system, a circuit board system, a game console system dedicated for entertainment purposes, an application-specific system. For example, the computer system 600 may take the form of a desktop computer, a laptop computer, a server, a workstation, a game console, an embedded system.
[0079] Furthermore, the computer system 600 may take the form of various other devices including, but not limited to a personal digital assistant (PDA) device, a mobile phone device, a smart phone, a television, and so forth. Additionally, although not shown, the computer system 600 may be coupled to a network (for example, a telecommunications network, a local area network (LAN), a wireless network, a wide area network (WAN) such as the Internet, a peer-to-peer network, a cable network, or the like) for communication purposes through an I / O interface 608
[0080] It should be understood that the arrangement of components illustrated in the figures described are exemplary and that other arrangements may be possible. It should also be understood that the various system components (and means) defined by the claims, described below, and illustrated in the various block diagrams represent components in some systems configured according to the subject matter disclosed herein. For example, one or more of these system components (and means) may be realized, in whole or in part, by at least some of the components illustrated in the arrangements illustrated in the described figures. In addition, while at least one of these components are implemented at least partially as an electronic hardware component, and therefore constitutes a machine, the other components may be implemented in software that when included in an execution environment constitutes a machine, hardware, or a combination of software and hardware.
[0081] Although the disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions, and alterations can be made herein without departing from the spirit and scope of the disclosure as defined by the appended claims.
Claims
CLAIMS1. A Content Distributed Network, CDN, controller (102) configured to minimize impact of one or more rate limiters, wherein the CDN controller (102) is configured to receive a first plurality of data packets (104A-N), each data packet comprising one or more packet features; generate a traffic histogram (108) for the packet feature based on the first plurality of data packets (104A-N), wherein the traffic histogram (108) for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the traffic histogram (108), and wherein the traffic histogram (108) indicates benign traffic quotas (BQ); determine that there is an attack undergoing; receive a second plurality of data packets (106A-N); determine an effect of a rate limiter on a packet feature by generating an attack histogram (110) for the packet feature based on the plurality of data packets, wherein the attack histogram (110) indicates attack traffic quotas (AQ); determining an Attack Differentiating Factor, ADF, for each bin, in the attack histogram (110) based on the formulaADF (f, i) = AQ(f, i) / BQ(f, i), wherein f indicates the packet feature and i indicates the histogram value bin, wherein the ADF indicates the effect of the rate limiter.
2. The CDN controller (102) according to claim 1, wherein the CDN controller (102) is further configured to determine the ADF based on the formula ADF (f, i) = AQ(f, i) / Max(BQ(f, i); MinBQ(f, i)), wherein MinBQ(f, i) is a set minimum value for the bin i for feature f during traffic prior to the attack.
3. The CDN controller (102) according to claim 2, wherein the CDN controller (102) is further configured to set the MinBQ to a default minimum value.
4. The CDN controller (102) according to any preceding claim, wherein the CDN controller (102) is further configured to select a rate limiter to be combined from a plurality of bin specificrate limiters, wherein each rate limiter is associated with one or more bin of appropriate feature’s histogram, wherein the controller is further configured to determine the ADF for one or more packet features, select the rate limiter to be enforced to be a rate limiter associated with the bin of packet feature having the highest ADF.
5. The CDN controller (102) according to claim 4, wherein the CDN controller (102) is further configured to determine that the there is a packet feature having an ADF exceeding a threshold value prior to selecting the rate limiter to be enforced.
6. The CDN controller according to claim 5, wherein the controller is further configured to determine that an ADF exceeds a threshold value utilizing a Kolmogorov-Smirnov Test.
7. The CDN controller (102) according to claim 5 or 6, wherein the CDN controller (102) is further configured to determine that there is no packet feature having an ADF exceeding a threshold value prior to selecting the rate limiter to be enforced, and in response thereto not selecting a rate limiter to be enforced.
8. The CDN controller (102) according to any of claims 4, 5, or 6, wherein the CDN controller (102) is further configured to determine if a further rate limiter to be enforced is to be selected by determining a traffic volume to be removed, RMV, (306) wherein the traffic volume to be removed is determined as the traffic volume exceeding an attack threshold value determine that the RMV (306) is above 0 and in response thereto select a further rate limiter based on the ADFs determined for the packet features.
9. The CDN controller (102) according to any of claims 4 to 8, wherein the CDN controller (102) is further configured to determine that the attack is still ongoing and in response thereto receive a further second plurality of data packets and determine a second attack histogram based on the further second plurality of data packets and ADFs for the packet features.
10. The CDN controller (102) according to claim 9, wherein the CDN controller (102) is further configured to determine that the attack is still ongoing at intervals.
11. The CDN controller (102) according to claim 10, wherein the intervals are at 45, 60, 75, 90, 105, 120, 135, 150, 165 or 180 seconds or any range there in-between.
12. The CDN controller (102) according to any of claims 4 to 11, wherein the CDN controller (102) is further configured to determine the RMV (306) as RMV = FF x SUM(AQ(f,i)), wherein FF is a Feedback Factor, determined as FFj = 1- RMVj-i / RMVj-2, wherein FFo = FFi= 1.
13. The CDN controller (102) according to any preceding claim, wherein the packet feature is a nbr packets, a nbr of open connections, a packet size, a header size, payload size, a protocol type, destination port, a source port, a fragment number, a Time to Live, TTL, a Transmission Control Protocol, TCP, flag, a sequence counter, or an Internet Protocol, IP identifier.
14. A method for use in a Content Distributed Network, CDN, controller (102) for minimizing impact of one or more rate limiters, the method comprising: receiving a first plurality of data packets (104A-N), each data packet comprising one or more packet features; generating a traffic histogram (108) for the packet feature based on the first plurality of data packets (104A-N), wherein the traffic histogram (108) for the packet feature is based on the packet feature matching a plurality of conditions, each condition corresponding to a bin in the traffic histogram (108), and wherein the traffic histogram (108) indicates benign traffic quotas (BQ); determining that there is an attack undergoing; receiving a second plurality of data packets (106A-N); determining an effect of a rate limiter on a packet feature by generating an attack histogram (110) for the packet feature based on the plurality of data packets, wherein the attack histogram (110) indicates attack traffic quotas (AQ), determining an Attack Differentiating Factor, ADF, for each bin, in the attack histogram (110) based on the formulaADF (f, i) = AQ(f, i) / BQ(f, i), wherein f indicates the packet feature and i indicates the histogram value bin, wherein the ADF indicates the effect of the rate limiter.
15. The method of claim 14, wherein the method is further for selecting a rate limiter to be combined from a plurality of bin specific rate limiters, wherein each rate limiter is associated with one or more bin of appropriate feature’s histogram, wherein the method further comprises determining the ADF for one or more packet features; selecting the rate limiter to be enforced to be a rate limiter associated with the bin of packet feature having the highest ADF.
16. A computer program product comprising program instructions for performing the method according to claim 14 or 15, when executed by one or more processors in a CDN system.
Citation Information
Patent Citations
Denial of service attacks characterization
US20030145232A1
Method and system for detecting and mitigating https flood attacks
US20200412750A1
