TA14-017A:UDP-Based Amplification Attacks   
      
   Original release date: January 17, 2014 | Last revised: April 13, 2016   
   Systems Affected   
   Certain application-layer protocols that rely on User Datagram Protocol (UDP)   
   have been identified as potential attack vectors:   
      
   DNS   
   NTP   
   SNMPv2   
   NetBIOS   
   SSDP   
   CharGEN   
   QOTD   
   BitTorrent   
   Kad   
   Quake Network Protocol   
   Steam Protocol   
   RIPv1   
   Multicast DNS (mDNS)   
   Portmap/RPC   
   Overview   
   A Distributed Reflective Denial of Service (DRDoS) attack is a form of   
   Distributed Denial of Service (DDoS) that relies on the use of publicly   
   accessible UDP servers, as well as bandwidth amplification factors, to   
   overwhelm a victim system with UDP traffic.   
      
   Description   
   UDP, by design, is a connection-less protocol that does not validate source IP   
   addresses. Unless the application-layer protocol uses countermeasures such as   
   session initiation in VOIP (voice over IP), it is very easy to forge the IP   
   packet datagram to include an arbitrary source IP address [1]. When many UDP   
   packets have their source IP address forged to the victim IP address, the   
   destination server (or amplifier) responds to the victim (instead of the   
   attacker), creating a reflected Denial of Service (DoS) Attack.   
      
   Recently, certain UDP protocols have been found to have particular responses   
   to certain commands that are much larger than the initial request. Previously,   
   attackers were limited linearly by the number of packets directly sent to the   
   target to conduct a DoS attack; now a single packet can generate tens or   
   hundreds of times the bandwidth in its response. This is called an   
   amplification attack, and when combined with a reflective DoS attack on a   
   large scale using multiple amplifiers and targeting a single victim, DDoS   
   attacks can be conducted with relative ease.   
      
   To measure the potential effect of an amplification attack, a metric called   
   the bandwidth amplification factor (BAF) is used. BAF can be calculated as the   
   number of UDP payload bytes that an amplifier sends to answer a request,   
   compared to the number of UDP payload bytes of the request [2] [3].   
      
   The list of known protocols—and their associated bandwidth amplification   
   factors—are listed below. US-CERT offers thanks to Christian Rossow for   
   providing this information. For more information on bandwidth amplification   
   factors, please see Christian's blog and associated research paper.   
      
   Protocol Bandwidth Amplification Factor Vulnerable Command   
   DNS 28 to 54 see: TA13-088A [4]   
   NTP 556.9 see: TA14-013A [5]   
   SNMPv2 6.3 GetBulk request   
   NetBIOS 3.8 Name resolution   
   SSDP 30.8 SEARCH request   
   CharGEN 358.8 Character generation request   
   QOTD 140.3 Quote request   
   BitTorrent 3.8 File search   
   Kad 16.3 Peer list exchange   
   Quake Network Protocol 63.9 Server info exchange   
   Steam Protocol 5.5 Server info exchange   
   Multicast DNS (mDNS) 2 to 10 Unicast query   
   RIPv1 131.24 Malformed request   
   Portmap (RPCbind) 7 to 28 Malformed request   
      
   In March 2015, Software Engineering Institute CERT issued Vulnerability Note   
   (VU#550620) describing the use of mDNS in DRDoS attacks. Attackers can   
   leverage mDNS by sending more information than can be handled by the device,   
   thereby causing a DoS. [6]   
      
   In July 2015, Akamai Technologies' Prolexic Security Engineering and Research   
   Team (PLXsert) issued a threat advisory describing a surge in DRDoS attacks   
   using the Routing Information Protocol version one (RIPv1). Malicious actors   
   are leveraging the behavior of RIPv1 for DDoS reflection through specially   
   crafted request queries [7].   
      
   In August 2015, Level 3 Threat Research Labs reported a new form of DRDoS   
   attack that uses portmap.  Attackers leverage the behavior of the portmap   
   service through spoofed requests and flood a victim’s network with UDP   
   traffic. [8]   
      
   Impact   
   Attackers can utilize the bandwidth and relative trust of large servers that   
   provide the above UDP protocols to flood victims with unwanted traffic, a DDoS   
   attack.   
      
   Solution   
   DETECTION   
   Detection of DRDoS attacks is not easy because of their use of large, trusted   
   servers that provide UDP services. Network operators of these exploitable   
   services may apply traditional DoS mitigation techniques. To detect a DRDOS   
   attack, watch out for abnormally large responses to a particular IP address,   
   which may indicate that an attacker is using the service.   
      
   If you are a victim of DRDoS attack, there are a few things you can do to   
   detect such activity and respond:   
      
   Detect and alert large UDP packets to higher order ports.   
   Detect and alert on any non-stateful UDP packets. (A simple snort example is   
   below. You will need to customize this approach to your environment with   
   whitelist and known services.) Simple Snort rule example for stateless UDP   
   check   
   var HOME_NET [10.10.10.20]   
   preprocessor stream5_global: track_ip yes, track_tcp yes,track_udp   
   yes,track_icmp no,max_tcp 262144, max_udp 131072   
   preprocessor stream5_ip: timeout 180   
   preprocessor stream5_tcp: policy first, use_static_footprint_sizes   
   preprocessor stream5_udp: timeout 180, ignore_any_rules   
   alert udp HOME_NET 1024: -> any any (msg:"UDP Session start"; fl   
   wbits:set,logged_in; flowbits:noalert; sid: 1001;)   
   alert udp any any -> HOME_NET 1024: (msg:"UDP Stateless"; flowbi   
   s:isnotset,logged_in; sid: 1002)   
      
   If you are an upstream provider maintain updated contacts and methods with   
   downstream customers to send alerts by network.   
   In general, network and server administrators for Internet service providers   
   (ISPs) should use the following best practices to avoid becoming amplifier   
   nodes:   
      
   Detect spoofed packets using network flow. (In order to validate before   
   blocking it, read more in the Mitigation section on blocking spoofed traffic.)   
   Monitor for an unusual number of requests to UDP services at risk using   
   network flow or other summarized network data.   
   Use network flow to detect service anomalies (bytes-per-packet,    
   ackets-per-second anomalies).   
   MITIGATION   
   If you are a victim of DRDoS attack there are a few things you can do to   
   mitigate this attack:   
      
   Use stateful UDP inspection to reduce impact on critical services on your   
   border firewall or border router (like reflexive ACL [9])   
   Using Border Gateway Protocol (BGP), create a Remotely Triggered Blackhole,   
   preferably in coordination with your upstream provider or ISP. [10]   
   Maintain a list of primary upstream provider emergency contacts to coordinate   
   response to the attack. If you are an upstream provider, conduct mitigation in   
   coordination with your downstream customers.   
   In general, network and server administrators for Internet service providers   
   (ISPs) should use the following as best practices to avoid becoming amplifier   
   nodes:   
      
   Keep your software and configuration up to date to deny or limit abuse (e.g.,   
   DNS response rate limit [11] [12] [13])   
   Disable and remove unwanted services or deny access to local services over the   
   Internet.   
   Enable network-based rate-limiting to legitimate services you provide over the   
   Internet using UDP-based protocols (e.g., quality of service (QoS) on   
   switching and routing devices)   
   Work with Customer Provider Edge manufactures for secure configuration and   
   software [14]   
   As a service provider, to avoid any misuse of Internet resources   
      
   Block spoofed packets by using ingress filtering (the Spoofer Project [15] and   
   IETF BCP 28 and BCP 39 guidelines [16])   
   Use traffic shaping on UDP service requests to ensure repeated access to   
   over-the-internet resources is not abusive.[17] [18]   
   References   
   [1] SIP: Session Initiation Protocol   
   [2] Amplification Hell: Abusing Network Protocols for DDoS   
   [3] Ampli?cation Hell: Revisiting Network Protocols for DDoS Abuse   
   [4] DNS Amplification Attacks   
   [5] NTP Amplification Attacks Using CVE-2013-5211   
   [6] VU#550620: Multicast DNS (mDNS) implementations may respond to unicast   
   queries originating outside the local link   
   [7] RIPv1 Reflection DDoS [Medium Risk]   
   [8] A New DDoS Reflection Attack: Portmapper; An Early Warning to the Industry   
   [9] Configuring IP Session Filtering (Reflexive Access Lists)   
   [10] Remotely-Triggered Black Hole (RTBH) Routing   
   [11] A Quick Introduction to Response Rate Limiting   
   [12] Network Ingress Filtering: Defeating Denial of Service Attacks Which   
   Employ IP Source Address Spoofing   
   [13] Ingress Filtering for Multihomed Networks   
   [14] Abuse of Customer Premise Equipment and Recommended Actions   
   [15] The Spoofer Project   
   [16] Abuse of Customer Premise Equipment and Recommended Actions   
   [17] An Architecture for Differentiated Services   
   [18] New Terminology and Clarifications for Diffserv   
   Revisions   
   February 9, 2014 – Initial Release   
   March 7, 2014 – Updated page to include research links   
   July 13, 2015 – Added RIPv1 as an attack vector   
   August 19, 2015 – Added Multicast DNS (mDNS) and Portmap (RPCbind) as attack   
   vectors   
   April 13, 2016 – Updated detection and mitigation information   
      
   ----------------------------------------------------------------   
   -------------- -   
   OTHER RESOURCES:   
   Contact Us | Security Publications | Alerts and Tips | Related Resources   
   STAY CONNECTED:   
   Sign up for email updates   
      
   SUBSCRIBER SERVICES:   
   Manage Preferences  |  Unsubscribe  |  Help   
      
      
   ----------------------------------------------------------------   
   -------------- -   
   This email was sent to Fido4cmech@lusfiber.net using GovDelivery, on behalf   
   of: United States Computer Emergency Readiness Team (US-CERT) · 245 Murray   
   Lane SW Bldg 410 · Washington, DC 20598 · (888) 282-0870 Powered by GovDelivery   
   === Cut ===   
      
   --   
   Keep the faith   :^)   
      
      Ben  aka cMech  Web: http|ftp|telnet://cmech.dynip.com   
                    Email: fido4cmech(at)lusfiber.net   
                 Home page: http://cmech.dynip.com/homepage/   
              WildCat! Board 24/7  +1-337-984-4794  any BAUD 8,N,1   
      
   --- GoldED+/W32-MSVC   
    * Origin: FIDONet - The Positronium Repository (1:393/68)