Elsevier

Computers & Security

Volume 109, October 2021, 102380
Computers & Security

A review of amplification-based distributed denial of service attacks and their mitigation

https://doi.org/10.1016/j.cose.2021.102380Get rights and content

Abstract

The rise of Distributed Denial of Service (DDoS) attacks have been steady in terms of the frequency and the impact of the attack. Traditionally, the attackers required control of a huge amount of resources to launch an attack. This has changed with the use of reflectors and amplifiers in DDoS attacks. A recent shift consisted of using other protocols than the traditional NTP and DNS protocols which were heavily used for ADDoS. In this paper, we review and organize amplification-based DDoS (ADDoS) attacks and associated countermeasures into a new taxonomy. Furthermore, we present a modus operandi of ADDoS attacks and analyze how it differs from traditional DDoS attacks. We also investigate how accessible ADDoS are for attackers with average resources. We survey readily available open-source scripts on GitHub and also the ADDoS features available in hire-to-DDoS platforms. We believe that accessibility and low-cost of hire-to-DDoS platforms are the major reasons for the increase of amplification-based DDoS attacks. Lastly, we provide a list of future directions that might be interesting for the community to focus on.

Introduction

The ability to perform distributed Denial of Service (DoS) and Distributed Denial of Service (DDoS) attacks to halt services on the Internet has been on the rise in recent years (Khalimonenko Alexander et al., 2018). Individuals with little effort and minimal cost can launch gigabytes of data using simple web interfaces provided by DDoS-for-hire services. To make attacks even easier, some of these services provide mobile applications to their customers (Google Play Store, 2018). Huge attacks launched by individuals with minimal technical knowledge and financial resources from user-friendly platforms is a current trend, according to Kaspersky (Makrushin Denis, 2017).

One of the recent and prominent methods of DDoS attacks is the Amplification-based DDoS (ADDoS). The recent 1.3 Tbps “memcached” reflection attack was an ADDoS example (Akamai, 2018). The size of the reflection attacks continues to grow due to the ability to use reflectors and amplifiers, resulting in more severe damage. The Network Security Labs at 360’s famous DDoSMon tool reported a surge of increase in memcached UDP amplifiers and reflector as shown in Fig. 1 (Yang, 2018).

In 2018, US based Security Company Cybereason reported that attackers embedded short ransom note and payment address details into the junk traffic of memcached servers (Krebsonsecurity, 2018). This makes it evident that the motive of amplification-based DDoS attacks range from taking down a competitor to severe crimes like extortion. The ransom note is shown in Fig. 2.

The impact of amplification-based DDoS attacks are even worse when the target resides on the Cloud. Economic Denial of Sustainability (EDoS), aka Fraudulent Resource Consumption (FRC), is a completely new type of DDoS attacks that targets cloud-hosted victims (Naresh Kumar, Sujatha, Kalva, Nagori, Katukojwala, Kumar, 2012, Vivinsandar, 2012). The Cloud’s resources are set to auto-scale up and down to manage cost effectively based on the requirements. DDoS attack’s potential effect would first and foremost lead to increasing usage of more resources in the Cloud (due to auto scaling), which in turn will increase the cost of the owner of the Cloud. Larger attacks can lead to a point of unavailability of services.

Modern botnets are not limited of personal computers, but include also handheld and IoT devices to increase their generated traffic (Akamai, 2021). Akamai released their report showing various headlines that caused serious disruption based on IoT devices. All of them such as Mirai, XOR and Spike were more than 300 Gbps, which indicates the trend in the volume of attack as shown in Fig. 3 (Akamai, 2021). We focus on understanding the major difference between a traditional and amplified DDoS attack in the next section.

Our contributions in this paper are as follows:

  • While several surveys reviewed DoS and DDoS attack and defense mechanisms, only few focused on Amplification-based DDoS (ADDoS). We provide an up-to-date survey of amplification-based DDoS attack.

  • We categorize notable ADDoS work and provide a new taxonomy for better understanding the subject and its’ complexities.

  • Using the existing literature, we derive a 4-step methodology that ADDoS attackers follow in order to perform ADDoS.

  • We survey readily available open source scripts that could be used to launch an ADDoS. Furthermore, we investigate some of the DDoS tools available online which have amplification-based attacks as a feature in them. Both the free scripts and the cheap DDoS tools (with amplification attack feature) sheds light into the ease of launching an ADDoS attack. We show that the ease of launching an ADDoS attack is a significant factor in the rise of of amplification-based attacks.

The rest of this paper is organized as follows: Section 2 reviews how traditional DDoS attacks are carried out and discusses how those attacks can be carried on or from the cloud. Section 3 focuses on understanding the fundamentals of ADDoS and how an attack is initiated by an attacker. Section 4 analyzes the work done in the area and provides a taxonomy of ADDoS based on attack and countermeasures. Section 5 reviews the accessibility of ADDoS in terms of availability of resource and the cost of carrying out an attack. Finally, Section 6 concludes the paper with the major research direction that we were able to derive.

Section snippets

Traditional DDoS and cloud

DDoS attacks were generally created using a large number of infected machines (zombies or bots). The attacker uses one platform to control all the machines which are directed generally using DyDNS or No-IP (Thomson Iain, 2014). The attacker typically spreads an infected executable to get a maximum number of bots. This group of bots is called a botnet. So the larger the botnet, the better the ability to create more damage. The botnet is controlled from the command and control platform which is

Amplification-based distributed denial of service (ADDoS)

From the traditional command and control style attack, DDoS attacks have evolved into amplification attacks. This has been the most common type of DDoS attack in recent years (Akamai, Musil Steven, Prince Matthew). There are two essential aspects to this type of attack:

  • 1.

    Reflection: The attacker spoofs the IP packets (IP spoofing) and changes the source address to the IP of the victim and sends it to some third party entity on the Internet.

  • 2.

    Amplification: The third-party entities (mostly

Taxonomy of ADDoS

It is important to have a taxonomy of the constantly changing ADDoS both from an attack and a mitigation perspective. This would help to better the work being done and to have a holistic image of the solutions.Colella and Colombini (2014) provided an attack taxonomy in 2014 and categorized the attack model which is very similar to our categorization for attack. Ryba et al. (2015) surveyed amplification area to a great extent covering a general understanding of amplification attacks and

Accessibility of ADDoS

We tried to analyze the ease of an attacker to create/launch an amplification-based DDoS attack based on two criteria:

  • 1.

    Availability of resources.

  • 2.

    Cost of the attack.

Conclusion and future research direction

We showed how reflection and amplification have changed the landscape of DDoS attacks. The major obstacle to releasing the full potential of DDoS attacks was the lack of resources available to attackers. Indeed, it required a huge botnet to launch an attack that could create some serious damage. However, with amplification-based DDoS attacks in play, lack of resources can be overcome by using these types of attacks. The impact of these attacks are on a rise and understanding how amplification

Declaration of Competing Interest

Authors declare that they have no conflict of interest.

Salih Ismail is a lecturer in the School of Science and Engineering in Curtin University, Dubai. He is currently pursuing his doctoral research in security of cloud computing. His research interests include Cyber Security, Internet of Things, Cloud Computing and Semantic Web. He graduated with a distinction in M.Sc. Computer Systems Management from Heriot-Watt University. He received his B.Sc. (hons) in Information Technology from University of Bedfordshire. Prior to academia he worked in the

References (109)

  • M. Anagnostopoulos et al.

    DNS amplification attack revisited

    Comput. Secur.

    (2013)
  • Apache Software Foundation, 2013. Chukwa - Apache....
  • M. Aupetit et al.

    Visualization of actionable knowledge to mitigate DrDoS attacks

    2016 IEEE Symposium on Visualization for Cyber Security (VizSec)

    (2016)
  • D. Beckett et al.

    HTTP/2 tsunami: investigating HTTP/2 proxy amplification DDoS attacks

    2017 Seventh International Conference on Emerging Security Technologies (EST)

    (2017)
  • K. Bhushan et al.

    Distributed denial of service (DDoS) attack mitigation in software defined network (SDN)-based cloud computing environment

    J. Ambient Intell. Humanized Comput.

    (2019)
  • E. Biagioni

    Preventing UDP Flooding Amplification Attacks with Weak Authentication

    2019 International Conference on Computing, Networking and Communications (ICNC)

    (2019)
  • E. Bohte et al.

    Evaluation of Current State of Amplification-based DDoS Attacks

    (2018)
  • Bombarding, 2016. GitHub - Bombarding/DNS-Amp-Attack: Perl Script for a DNS Amplification Attack....
  • P. Bosshart et al.

    P4: programming protocol-independent packet processors

    ACM SIGCOMM Comput. Commun. Rev.

    (2014)
  • T. Böttger et al.

    DoS amplification attacks–protocol-agnostic detection of service abuse in amplifier networks

    International Workshop on Traffic Monitoring and Analysis

    (2015)
  • L. Cai et al.

    A behavior-based method for detecting DNS amplification attacks

    2016 10th International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS)

    (2016)
  • Censys, 2021. Home - Censys....
  • Chadillac, 2017. GitHub - chadillac/mdns_recon....
  • Chang Liu et al.

    Detect the reflection amplification attack based on UDP protocol

    2015 10th International Conference on Communications and Networking in China (ChinaCom)

    (2015)
  • C.-C. Chen et al.

    Detecting amplification attacks with software defined networking

    2017 IEEE Conference on Dependable and Secure Computing

    (2017)
  • S.-J. Choi et al.

    A study on reduction of DDoS amplification attacks in the UDP-based CLDAP protocol

    2017 4th International Conference on Computer Applications and Information Processing Technology (CAIPT)

    (2017)
  • A. Colella et al.

    Amplification DDoS attacks: emerging threats and defense strategies

    International Conference on Availability, Reliability, and Security

    (2014)
  • J. Czyz et al.

    Taming the 800 pound gorilla

    Proceedings of the 2014 Conference on Internet Measurement Conference - IMC ’14

    (2014)
  • A. Dahiya et al.

    Multi attribute auction based incentivized solution against DDoS attacks

    Comput. Secur.

    (2020)
  • David Wittman, 2016. GitHub - DavidWittman/wpxmlrpcbrute: Brute force WordPress sites vulnerabile to XML-RPC...
  • David9705, 2019. GitHub - david9705/DNS-Amplification-Attack....
  • R.V. Deshmukh et al.

    Understanding DDoS Attack & its Effect in Cloud Environment

    Procedia Comput. Sci.

    (2015)
  • DrizzleRisk, 2018. GitHub - DrizzleRisk/NTPDoser: NTP Doser is a NTP Amplification DoS/DDoS attack tool for...
  • Z. Durumeric et al.

    ZMap: fast internet-wide scanning and its security applications

    USENIX Secur. Symp.

    (2013)
  • Ethanwilloner, 2015. GitHub - ethanwilloner/DNS-Amplification-Attack....
  • Euphie, 2018. GitHub - Euphie/dns-amplification-attack....
  • C. Fachkha et al.

    Fingerprinting internet DNS amplification DDoSactivities

    2014 6th International Conference on New Technologies, Mobility and Security (NTMS)

    (2014)
  • Fred Klassen, 2014. tcprewrite....
  • Y. Gao et al.

    A machine learning based approach for detecting DrDoSattacks and its performance evaluation

    2016 11th Asia Joint Conference on Information Security (AsiaJCIS)

    (2016)
  • J.J. Gondim et al.

    Mirror saturation in amplified reflection distributed denial of service: a case of study using SNMP, SSDP, NTP and DNS protocols

    Future Gener. Comput. Syst.

    (2020)
  • Google Play Store, 2018. Android Apps on Google Play....
  • V. Gupta et al.

    Mitigating DNS amplification attacks using a set of geographically distributed SDN routers

    2018 International Conference on Advances in Computing, Communications and Informatics (ICACCI)

    (2018)
  • Hacker900123, 2019. GitHub - hacker900123/Amplifier: A Python2 Script Made To DDoS A Website Using Ntp, Dns, SSDP,...
  • N. Hoque et al.

    Botnet in DDoS attacks: trends and challenges

    IEEE Commun. Surv. Tutor.

    (2015)
  • Y. Huang et al.

    Thinoram: towards practical oblivious data access in fog computing environment

    IEEE Trans. Serv. Comput.

    (2019)
  • D. Huistra

    Detecting reflection attacks in DNS flows

    19th Twente Student Conference on IT

    (2013)
  • Hyperbolas, 2019. GitHub - Hyperbolas/DNS-Amplification-Attack: My NCTU tiny project....
  • X. Jing et al.

    A reversible sketch-based method for detecting and mitigating amplification attacks

    J. Netw. Comput. Appl.

    (2019)
  • A.S. Jose et al.

    Automatic detection and rectification of DNS reflection amplification attacks with hadoop MapReduce and Chukwa

    2014 Fourth International Conference on Advances in Computing and Communications

    (2014)
  • M. Karami et al.

    Stress Testing the Booters: Understanding and Undermining the Business of DDoS Services

    (2015)
  • Cited by (12)

    • Large-scale empirical evaluation of DNS and SSDP amplification attacks

      2022, Journal of Information Security and Applications
    • Energy efficiency in 5G systems: A systematic literature review

      2024, International Journal of Knowledge-Based and Intelligent Engineering Systems
    • A novel sensitive DDoS attacks against statistical test in network traffic fusion

      2023, Transactions on Emerging Telecommunications Technologies
    View all citing articles on Scopus

    Salih Ismail is a lecturer in the School of Science and Engineering in Curtin University, Dubai. He is currently pursuing his doctoral research in security of cloud computing. His research interests include Cyber Security, Internet of Things, Cloud Computing and Semantic Web. He graduated with a distinction in M.Sc. Computer Systems Management from Heriot-Watt University. He received his B.Sc. (hons) in Information Technology from University of Bedfordshire. Prior to academia he worked in the industry in various posts like IT Manager, Technology Officer, and Database Administrator etc. and sat on many advisory boards for tech start-ups. He is currently pursuing his Ph.D. from Heriot-Watt University on Security of Cloud Computing.

    Hani Ragab Hassen received his Ph.D. from the University of Technology of Compiegne, France in 2007. He was a lecturer in Information Security at the University of Kent, UK from 2007 to 2012. He is currently an associate professor with Heriot-Watt University. His main research interests include malware analysis, ethical hacking, peer-to-peer systems and cryptographic key management.

    Mike Just earned his Ph.D. in Computer Science from Carleton University (Ottawa, Canada) in 1998 under the direction of Paul Van Oorschot and Evangelos Kranakis. From 1998 till 2002 he was a Research Scientist at Entrust, a security software company, where he performed code security reviews, contributed to new product developments and created new IETF and W3C standards, primarily related to authentication and key management. From 2002 till 2008 he worked for the Government of Canada. In 2003, he designed the “account recovery” interface for the Government Online portal, used by more than six million citizens. I also contributed to design of the portal’s “epass” authentication service. In 2008 he made the leap to full-time academic work, where he continue to build on my usable security research, as well as applying machine learning to security. He started as a Research Fellow at the University of Edinburgh in 2008, then in 2010 he became a Lecturer and then Senior Lecturer at Glasgow Caledonian University, and in 2015 he moved to Heriot-Watt University as an Associate Professor.

    Dr. Hind Zantout MBCS SFHEA is the joint Head of the Computer Science Department (Teaching) at Heriot-Watt University and Programme Director of M.Sc. Data Science at the Dubai campus. Dubai. Her academic and research interests include semantic technologies, machine learning and interdisciplinary topics that link Computing and Informatics with Business and Society. Under this umbrella, Big Data, Innovation, Data Mining and Knowledge Management are employed to support work on Healthcare, the Knowledge Economy and Smart Communities.

    View full text