Denial-of-service attack

In computing, a denial-of-service attack (DoS attack) is a cyber-attack in which the perpetrator seeks to make a machine or network resource unavailable to its intended users by temporarily or indefinitely disrupting services of a host connected to a network. Denial of service is typically accomplished by flooding the targeted machine or resource with superfluous requests in an attempt to overload systems and prevent some or all legitimate requests from being fulfilled.

In a distributed denial-of-service attack (DDoS attack), the incoming traffic flooding the victim originates from many different sources. More sophisticated strategies are required to mitigate this type of attack, as simply attempting to block a single source is insufficient because there are multiple sources.

A DoS or DDoS attack is analogous to a group of people crowding the entry door of a shop, making it hard for legitimate customers to enter, thus disrupting trade.

Criminal perpetrators of DoS attacks often target sites or services hosted on high-profile web servers such as banks or credit card payment gateways. Revenge, blackmail and hacktivism can motivate these attacks.

History

Panix, the third-oldest ISP in the world, was the target of what is thought to be the first DoS attack. On September 6, 1996, Panix was subject to a SYN flood attack, which brought down its services for several days while hardware vendors, notably Cisco, figured out a proper defense.

Another early demonstration of the DoS attack was made by Khan C. Smith in 1997 during a DEF CON event, disrupting Internet access to the Las Vegas Strip for over an hour. The release of sample code during the event led to the online attack of Sprint, EarthLink, E-Trade and other major corporations in the year to follow.

In September 2017, Google Cloud experienced an attack with a peak volume of . On March 5, 2018, an unnamed customer of the US-based service provider Arbor Networks fell victim to the largest DDoS to that date, reaching a peak of about . The previous record had been set a few days earlier, on March 1, 2018, when GitHub was hit by an attack of .

In February 2020, Amazon Web Services experienced an attack with a peak volume of . In July 2021, CDN Provider Cloudflare boasted of protecting its client from a DDoS attack from a global Mirai botnet that was up to 17.2 million requests per second. Russian DDoS Prevention provider Yandex said it blocked a HTTP pipelining DDoS attack on Sept. 5. 2021 that originated from unpatched Mikrotik networking gear.

Types

Denial-of-service attacks are characterized by an explicit attempt by attackers to prevent legitimate use of a service. There are two general forms of DoS attacks: those that crash services and those that flood services. The most serious attacks are distributed.

A distributed denial-of-service (DDoS) attack occurs when multiple systems flood the bandwidth or resources of a targeted system, usually one or more web servers. A DDoS attack uses more than one unique IP address or machines, often from thousands of hosts infected with malware. A distributed denial of service attack typically involves more than around 3–5 nodes on different networks; fewer nodes may qualify as a DoS attack but is not a DDoS attack.

Multiple machines can generate more attack traffic than one machine, multiple attack machines are harder to turn off than one attack machine, and the behavior of each attack machine can be stealthier, making it harder to track and shut down. Since the incoming traffic flooding the victim originates from different sources, it may be impossible to stop the attack simply by using ingress filtering. It also makes it difficult to distinguish legitimate user traffic from attack traffic when spread across multiple points of origin. As an alternative or augmentation of a DDoS, attacks may involve forging of IP sender addresses ( IP address spoofing) further complicating identifying and defeating the attack. These attacker advantages cause challenges for defense mechanisms. For example, merely purchasing more incoming bandwidth than the current volume of the attack might not help, because the attacker might be able to simply add more attack machines.

The scale of DDoS attacks has continued to rise over recent years, by 2016 exceeding a terabit per second. Some common examples of DDoS attacks are UDP flooding, SYN flooding and DNS amplification.

Yo-yo attack

A yo-yo attack is a specific type of DoS/DDoS aimed at cloud-hosted applications which use autoscaling. The attacker generates a flood of traffic until a cloud-hosted service scales outwards to handle the increase of traffic, then halts the attack, leaving the victim with over-provisioned resources. When the victim scales back down, the attack resumes, causing resources to scale back up again. This can result in a reduced quality of service during the periods of scaling up and down and a financial drain on resources during periods of over-provisioning while operating with a lower cost for an attacker compared to a normal DDoS attack, as it only needs to be generating traffic for a portion of the attack period.

Application layer attacks

An application layer DDoS attack (sometimes referred to as layer 7 DDoS attack) is a form of DDoS attack where attackers target application-layer processes. The attack over-exercises specific functions or features of a website with the intention to disable those functions or features. This application-layer attack is different from an entire network attack, and is often used against financial institutions to distract IT and security personnel from security breaches. In 2013, application-layer DDoS attacks represented 20% of all DDoS attacks. According to research by Akamai Technologies, there have been "51 percent more application layer attacks" from Q4 2013 to Q4 2014 and "16 percent more" from Q3 2014 to Q4 2014. In November 2017; Junade Ali, an engineer at Cloudflare noted that whilst network-level attacks continue to be of high capacity, they were occurring less frequently. Ali further noted that although network-level attacks were becoming less frequent, data from Cloudflare demonstrated that application-layer attacks were still showing no sign of slowing down. In December 2021, following the Log4Shell security vulnerability, a second vulnerability in the open source Log4j library was discovered which could lead to application layer DDoS attacks.

Application layer

The OSI model (ISO/IEC 7498-1) is a conceptual model that characterizes and standardizes the internal functions of a communication system by partitioning it into abstraction layers. The model is a product of the Open Systems Interconnection project at the International Organization for Standardization (ISO). The model groups similar communication functions into one of seven logical layers. A layer serves the layer above it and is served by the layer below it. For example, a layer that provides error-free communications across a network provides the communications path needed by applications above it, while it calls the next lower layer to send and receive packets that traverse that path.

In the OSI model, the definition of its application layer is narrower in scope than is often implemented. The OSI model defines the application layer as being the user interface. The OSI application layer is responsible for displaying data and images to the user in a human-recognizable format and to interface with the presentation layer below it. In an implementation, the application and presentation layers are frequently combined.

Method of attack

The simplest DoS attack relies primarily on brute force, flooding the target with an overwhelming flux of packets, oversaturating its connection bandwidth or depleting the target's system resources. Bandwidth-saturating floods rely on the attacker's ability to generate the overwhelming flux of packets. A common way of achieving this today is via distributed denial-of-service, employing a botnet.

An application layer DDoS attack is done mainly for specific targeted purposes, including disrupting transactions and access to databases. It requires fewer resources than network layer attacks but often accompanies them. An attack may be disguised to look like legitimate traffic, except it targets specific application packets or functions. The attack on the application layer can disrupt services such as the retrieval of information or search functions on a website.

Advanced persistent DoS

An advanced persistent DoS (APDoS) is associated with an advanced persistent threat and requires specialized DDoS mitigation. These attacks can persist for weeks; the longest continuous period noted so far lasted 38 days. This attack involved approximately 50+ petabits (50,000+ terabits) of malicious traffic.

Attackers in this scenario may tactically switch between several targets to create a diversion to evade defensive DDoS countermeasures but all the while eventually concentrating the main thrust of the attack onto a single victim. In this scenario, attackers with continuous access to several very powerful network resources are capable of sustaining a prolonged campaign generating enormous levels of un-amplified DDoS traffic.

APDoS attacks are characterized by:
* advanced reconnaissance (pre-attack OSINT and extensive decoyed scanning crafted to evade detection over long periods)
* tactical execution (attack with both primary and secondary victims but the focus is on primary)
* explicit motivation (a calculated end game/goal target)
* large computing capacity (access to substantial computer power and network bandwidth)
* simultaneous multi-threaded OSI layer attacks (sophisticated tools operating at layers 3 through 7)
* persistence over extended periods (combining all the above into a concerted, well-managed attack across a range of targets).

Denial-of-service as a service

Some vendors provide so-called ''booter'' or ''stresser'' services, which have simple web-based front ends, and accept payment over the web. Marketed and promoted as stress-testing tools, they can be used to perform unauthorized denial-of-service attacks, and allow technically unsophisticated attackers access to sophisticated attack tools. Usually powered by a botnet, the traffic produced by a consumer stresser can range anywhere from 5-50 Gbit/s, which can, in most cases, deny the average home user internet access.

Symptoms

The United States Computer Emergency Readiness Team (US-CERT) has identified symptoms of a denial-of-service attack to include:
* unusually slow network performance (opening files or accessing websites),
* unavailability of a particular website, or
* inability to access any website.

Attack techniques

Attack tools

In cases such as MyDoom and Slowloris, the tools are embedded in malware and launch their attacks without the knowledge of the system owner. Stacheldraht is a classic example of a DDoS tool. It uses a layered structure where the attacker uses a client program to connect to handlers which are compromised systems that issue commands to the zombie agents which in turn facilitate the DDoS attack. Agents are compromised via the handlers by the attacker using automated routines to exploit vulnerabilities in programs that accept remote connections running on the targeted remote hosts. Each handler can control up to a thousand agents.

In other cases a machine may become part of a DDoS attack with the owner's consent, for example, in Operation Payback organized by the group Anonymous. The Low Orbit Ion Cannon has typically been used in this way. Along with High Orbit Ion Cannon a wide variety of DDoS tools are available today, including paid and free versions, with different features available. There is an underground market for these in hacker-related forums and IRC channels.

Application-layer attacks

Application-layer attacks employ DoS-causing exploits and can cause server-running software to fill the disk space or consume all available memory or CPU time. Attacks may use specific packet types or connection requests to saturate finite resources by, for example, occupying the maximum number of open connections or filling the victim's disk space with logs. An attacker with shell-level access to a victim's computer may slow it until it is unusable or crash it by using a fork bomb. Another kind of application-level DoS attack is XDoS (or XML DoS) which can be controlled by modern web application firewalls (WAFs).

All attacks belonging to the category of ''timeout exploiting'' Slow DoS Attacks implement an application-layer attack. Examples of threats are Slowloris, establishing pending connections with the victim, or SlowDroid, an attack running on mobile devices.

Another target of DDoS attacks may be to produce added costs for the application operator, when the latter uses resources based on cloud computing. In this case, normally application-used resources are tied to a needed quality of service (QoS) level (e.g. responses should be less than 200 ms) and this rule is usually linked to automated software (e.g. Amazon CloudWatch) to raise more virtual resources from the provider to meet the defined QoS levels for the increased requests. The main incentive behind such attacks may be to drive the application owner to raise the elasticity levels to handle the increased application traffic, to cause financial losses, or force them to become less competitive.

A ''banana attack'' is another particular type of DoS. It involves redirecting outgoing messages from the client back onto the client, preventing outside access, as well as flooding the client with the sent packets. A LAND attack is of this type.

Degradation-of-service attacks

Pulsing zombies are compromised computers that are directed to launch intermittent and short-lived floodings of victim websites with the intent of merely slowing it rather than crashing it. This type of attack, referred to as ''degradation-of-service'', can be more difficult to detect and can disrupt and hamper connection to websites for prolonged periods of time, potentially causing more overall disruption than a denial-of-service attack. Exposure of degradation-of-service attacks is complicated further by the matter of discerning whether the server is really being attacked or is experiencing higher than normal legitimate traffic loads.

Distributed DoS attack

If an attacker mounts an attack from a single host, it would be classified as a DoS attack. Any attack against availability would be classed as a denial-of-service attack. On the other hand, if an attacker uses many systems to simultaneously launch attacks against a remote host, this would be classified as a DDoS attack.

Malware can carry DDoS attack mechanisms; one of the better-known examples of this was MyDoom. Its DoS mechanism was triggered on a specific date and time. This type of DDoS involved hardcoding the target IP address before releasing the malware and no further interaction was necessary to launch the attack.

A system may also be compromised with a trojan containing a zombie agent. Attackers can also break into systems using automated tools that exploit flaws in programs that listen for connections from remote hosts. This scenario primarily concerns systems acting as servers on the web. Stacheldraht is a classic example of a DDoS tool. It uses a layered structure where the attacker uses a client program to connect to handlers, which are compromised systems that issue commands to the zombie agents, which in turn facilitate the DDoS attack. Agents are compromised via the handlers by the attacker. Each handler can control up to a thousand agents. In some cases a machine may become part of a DDoS attack with the owner's consent, for example, in Operation Payback, organized by the group Anonymous. These attacks can use different types of internet packets such as TCP, UDP, ICMP, etc.

These collections of compromised systems are known as botnets. DDoS tools like Stacheldraht still use classic DoS attack methods centered on IP spoofing and amplification like smurf attacks and fraggle attacks (types of bandwidth consumption attacks). SYN floods (a resource starvation attack) may also be used. Newer tools can use DNS servers for DoS purposes. Unlike MyDoom's DDoS mechanism, botnets can be turned against any IP address. Script kiddies use them to deny the availability of well known websites to legitimate users. More sophisticated attackers use DDoS tools for the purposes of extortionincluding against their business rivals.

It has been reported that there are new attacks from internet of things (IoT) devices that have been involved in denial of service attacks. In one noted attack that was made peaked at around 20,000 requests per second which came from around 900 CCTV cameras.

UK's GCHQ has tools built for DDoS, named PREDATORS FACE and ROLLING THUNDER.

Simple attacks such as SYN floods may appear with a wide range of source IP addresses, giving the appearance of a distributed DoS. These flood attacks do not require completion of the TCP three-way handshake and attempt to exhaust the destination SYN queue or the server bandwidth. Because the source IP addresses can be trivially spoofed, an attack could come from a limited set of sources, or may even originate from a single host. Stack enhancements such as SYN cookies may be effective mitigation against SYN queue flooding but do not address bandwidth exhaustion.

DDoS extortion

In 2015, DDoS botnets such as DD4BC grew in prominence, taking aim at financial institutions. Cyber-extortionists typically begin with a low-level attack and a warning that a larger attack will be carried out if a ransom is not paid in bitcoin. Security experts recommend targeted websites to not pay the ransom. The attackers tend to get into an extended extortion scheme once they recognize that the target is ready to pay.

HTTP slow POST DoS attack

First discovered in 2009, the HTTP slow POST attack sends a complete, legitimate HTTP POST header, which includes a ''Content-Length'' field to specify the size of the message body to follow. However, the attacker then proceeds to send the actual message body at an extremely slow rate (e.g. 1 byte/110 seconds). Due to the entire message being correct and complete, the target server will attempt to obey the ''Content-Length'' field in the header, and wait for the entire body of the message to be transmitted, which can take a very long time. The attacker establishes hundreds or even thousands of such connections until all resources for incoming connections on the victim server are exhausted, making any further connections impossible until all data has been sent. It is notable that unlike many other DDoS or DDoS attacks, which try to subdue the server by overloading its network or CPU, an HTTP slow POST attack targets the ''logical'' resources of the victim, which means the victim would still have enough network bandwidth and processing power to operate. Combined with the fact that the Apache HTTP Server will, by default, accept requests up to 2GB in size, this attack can be particularly powerful. HTTP slow POST attacks are difficult to differentiate from legitimate connections and are therefore able to bypass some protection systems. OWASP, an open source web application security project, released a tool to test the security of servers against this type of attack.

Challenge Collapsar (CC) attack

A Challenge Collapsar (CC) attack is an attack where standard HTTP requests are sent to a targeted web server frequently. The Uniform Resource Identifiers (URIs) in the requests require complicated time-consuming algorithms or database operations which may exhaust the resources of the targeted web server.

In 2004, a Chinese hacker nicknamed KiKi invented a hacking tool to send these kinds of requests to attack a NSFOCUS firewall named ''Collapsar'', and thus the hacking tool was known as ''Challenge Collapsar'', or ''CC'' for short. Consequently, this type of attack got the name ''CC attack''.

Internet Control Message Protocol (ICMP) flood

A smurf attack relies on misconfigured network devices that allow packets to be sent to all computer hosts on a particular network via the broadcast address of the network, rather than a specific machine. The attacker will send large numbers of IP packets with the source address faked to appear to be the address of the victim. Most devices on a network will, by default, respond to this by sending a reply to the source IP address. If the number of machines on the network that receive and respond to these packets is very large, the victim's computer will be flooded with traffic. This overloads the victim's computer and can even make it unusable during such an attack.

Ping flood is based on sending the victim an overwhelming number of ping packets, usually using the ''ping'' command from Unix-like hosts. It is very simple to launch, the primary requirement being access to greater bandwidth than the victim.

Ping of death is based on sending the victim a malformed ping packet, which will lead to a system crash on a vulnerable system.

The BlackNurse attack is an example of an attack taking advantage of the required Destination Port Unreachable ICMP packets.

Nuke

A Nuke is an old-fashioned denial-of-service attack against computer networks consisting of fragmented or otherwise invalid ICMP packets sent to the target, achieved by using a modified ping utility to repeatedly send this corrupt data, thus slowing down the affected computer until it comes to a complete stop.

A specific example of a nuke attack that gained some prominence is the WinNuke, which exploited the vulnerability in the NetBIOS handler in Windows 95