| DOTS | Y. Hayashi |
| Internet-Draft | NTT |
| Intended status: Informational | K. Nishizuka |
| Expires: January 21, 2020 | NTT Communications |
| M. Boucadair | |
| Orange | |
| July 20, 2019 |
DDoS Mitigation Offload: DOTS Applicability and Deployment Considerations
draft-hayashi-dots-dms-offload-00
This document describes a deployment scenario to assess the applicability of DOTS protocols together with a discussion on DOTS deployment considerations of such scenario. This scenario assumes that a DMS (DDoS Mitigation System) whose utilization rate is high sends its blocked traffic information to an orchestrator using DOTS protocols, then the orchestrator requests forwarding nodes such as routers to filter the traffic. Doing so enables service providers to mitigate the DDoS attack traffic automatically while ensuring interoperability and distributed filter enforcement.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on January 21, 2020.
Copyright (c) 2019 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Volume-based distributed Denial-of-Service (DDoS) attacks such as DNS amplification attacks are critical threats to be handled by service providers. When such attacks occur, service providers have to mitigate them immediately to protect or recover their services.
Therefore, for the service providers to immediately protect their network services from DDoS attacks, DDoS mitigation needs to be automated. To automate DDoS attack mitigation, it is desirable that multi-vendor elements involved in DDoS attack detection and mitigation collaborate and support standard interfaces to communicate.
DDoS Open Threat Signaling (DOTS) is a set of protocols for real-time signaling, threat-handling requests, and data filtering between the multi-vendor elements [I-D.ietf-dots-signal-channel] [I-D.ietf-dots-signal-call-home] [I-D.ietf-dots-signal-filter-control] [I-D.ietf-dots-data-channel]. This document describes an automated DDoS Mitigation offload scenario inherited from the DDoS orchestration scenario [I-D.ietf-dots-use-cases], which ambitions to enable cost-effective DDoS Mitigation. Furthermore, this document describes deployment consideration for network operators who carry out this scenario using DOTS protocols in their network.
This document aims to assess to what extent DOTS protocols can be used to provide the intended functionality and identify any gaps.
The readers should be familiar with the terms defined in [RFC8612] [I-D.ietf-dots-use-cases]
In addition, this document makes use of the following terms:
In general, DDoS countermeasures are divided into detection and filtering. Detection is technically challenging given the dynamic of attacks and sophisticated attack strategies. DDoS Mitigation System (DMS) can detect attack traffic based on a specific technology (provided and supposed to be updated and maintained by vendors to detect complex attacks), so service providers can increase DDoS countermeasure level by deploying the DMS in their network.
However, the number/capacity of DMS instances that can be deployed in a service providers network is limited due to equipment cost and dimensioning matters. Thus, DMS's utilization rate can reach its maximum capacity faster when the volume of DDoS attacks is enormous. When the rate reaches maximum capacity, the mitigation strategy needs to offload mitigation actions from the DMS to cost-effective forwarding nodes such as routers.
This section describes offloading mitigation actions from DMS whose utilization rate is high to cost-effective forwarding node using DOTS protocols. This section does not consider deployments where the network orchestrator and DMS are co-located.
Figures 1 and 2 show a sample component diagram and a sequence diagram of the deployment scenario, respectively.
+--------------+ +-----------+
| | | DDoS |+
| Orchestrator |<-------| mitigation||
| |S DOTS C| systems ||
+--------------+ +-----------+|
| +----------+
| e.g., BGP, BGP Flowspec
|
| +------------------+
+->| Forwarding nodes |+
+------------------+|
+-----------------+
* C is for DOTS Client function
* S is for DOTS Server function
Figure 1: Component Diagram of DDoS Mitigation Offload Scenario
The component diagram shown in Figure 1 differs from that of DDoS Orchestration scenario in [I-D.ietf-dots-use-cases] in some respects. First, the DMS embeds a DOTS client to send DOTS requests to the orchestrator. Second, the orchestrator sends a request to underlying forwarding nodes to filter the attack traffic.
+------------+ +----------+ +------------+
| | |DDoS |+ | Forwarding |+
|Orchestrator| |Mitigation|| | Nodes ||
| | |Systems || | ||
+------------+ +----------+| +------------+|
| +----------+ +------------+
| | |
| DOTS Request | |
|S<----------------------C| |
| | |
| e.g., BGP, BGP Flowspec | |
| Filter Attack Traffic | |
|-------------------------|------------->|
| | |
* C is for DOTS Client function
* S is for DOTS Server function
Figure 2: Sequence Diagram of DDoS Mitigation Offload Scenario
In this scenario, it is assumed that volume based attack already hits a network and attack traffic is detected and blocked by a DMS in the network. When the volume-based attack becomes intense, DMS's utilization rate can reach a certain threshold (e.g., maximum capacity). Then, the DMS sends a DOTS request as offload request to the orchestrator with the actions to enforce on the traffic. After that, the orchestrator requests the forwarding nodes to filter attack traffic by dissemination of flow specification rules protocols such as BGP Flowspec [RFC5575] on the basis of the blocked traffic information.
This schenario is divided into two cases based on type of link between the DMS and the orchestrator: "out-of-band case" and "in-band case".
"Out-of-band case" is that the DMS sends a DOTS request to the orchestrator with blocked traffic information by the DMS via out-of-band link. The link is not congested when it is under volume attack-time, so the link can convey a lot of information.
On the other hand, "in-band case" is that the DMS sends a mitigation request to the orchestrator with blocked traffic information by the DMS via in-band channel. The link can be congested when it is under volume attack-time, so the link can convey limited information.
This section describes deployment considerations: what type of DOTS protocol can be used and what type of information can be conveyed and effective by DOTS protocol in this scenario. Figure 3 shows overview of the DOTS signaling method and conveyed information for the out-of-band case and in-band case.
The volume of information should be considered carefully when DOTS protocol is used in the in-band case. What type of information can be conveyed by DMS relays on attack type detected by the DMS: reflection attack or non-reflection attack. When it is under non-reflection attack, src_ip and src_port information cannot be conveyed because attackers usually randomize the parameters so number of its become enormous. On the other hand, when it is under reflection attack, dst_port information cannot be conveyed because attackers usually randomize src_port so the number of dst_port of attack packets reached to victim become enormous. Furthermore, when it is under reflection attack, src_ip information cannot be conveyed when number of reflector is enormous.
+-------------+-----------------------------------+------------------+
| | Reflection Attack | Non-Reflection |
| | | Attack |
+-------------+-----------------------------------+------------------+
| Out-of-band | Attack Time |
| case | Method : Data Channel |
| | Info : src_ip, src_port, dst_ip, dst_port, protocol |
+-------------+-----------------------------------+------------------+
| In-band | Attack Time | Attack Time |
| case | (Number of reflector is small) | Method : Signal |
| | Method : Signal Channel with src | Channel |
| | Info : src_ip, src_port, | Info : dst_ip, |
| | dst_ip, protocol | dst_port, |
| +-----------------------------------+ protocol |
| | Attack Time | |
| | (Number of reflector is enormous) | |
| | Method : Signal Channel with src | |
| | Info : src_port, dst_ip, protocol | |
| +-----------------------------------+------------------+
| | Peace Time | Peace Time |
| | Method : Data Channel | Method : Data |
| | Info : src_port, | Channel |
| | dst_ip, protocol | Info : dst_ip, |
| | | dst_port, |
| | | protocol |
| | | |
| | Attack Time | Attack Time |
| | Method : Signal Channel | Method : Signal |
| | Control Filtering | Channel |
| | Info : ACL name | Control Filtering|
| | | Info : ACL name |
|-------------+------------------------------------------------------+
Figure 3: Signaling Method and Conveyed Information
About offloading DMS against reflection attack, the current signal channel [I-D.ietf-dots-signal-channel] is insufficient in terms of conveying source information. On the other hand, both signal channel extensions defined in [I-D.ietf-dots-signal-call-home] (called, source-* clauses hereafter) and the filtering control extensions [I-D.ietf-dots-signal-filter-control] allow for sending source information.
Using src-* attributes defined in [I-D.ietf-dots-signal-call-home] enables signal channel to convey src_ip information and src_port information in attack time. On the other hand, filtering control extensions can activate filtering rule configured in peacetime. Filtering rule for well-known port numbers abused for reflection attack can be configured to DOTS server in peacetime. However, filtering rule for reflector's IP address in attack time can't be known in peace time. So filtering control expansion can convey src_port information but can't send src_ip information against reflection attack. About sending source information in the DMS offload s scenario, the capability of the call home extension encompasses the capabilities of the filtering control extension.
The following sub-sections describes example of use DOTS protocols in each case.
In this case, the link is not congested when it is under volume attack-time, so DOTS data channel [I-D.ietf-dots-data-channel] is suitable because DOTS data channel has capability of conveying the drop-listed filtering rules including (src_ip, src_port, dst_ip, dst_port, protocol) information (and other actions such as 'rate-limit').
The procedure to use DOTS Data Channel in such case is as follows:
{
"ietf-dots-data-channel:acls": {
"acl": [
{
"name": "DMS_Offload_scenario_ACL",
"type": "ipv4-acl-type",
"activation-type": "immediate",
"aces": {
"ace": [
{
"name": "DMS_Offload_scenario_ACE_00",
"matches": {
"ipv4": {
"destination-ipv4-network": "192.0.2.2/32",
"source-ipv4-network": "203.0.113.2/32",
"protocol":17
},
"udp": {
"source-port": {
"operator": "eq",
"port": 53
}
}
},
"actions": {
"forwarding": "drop"
}
},
{
"name": "DMS_Offload_scenario_ACE_01",
"matches": {
"ipv4": {
"destination-ipv4-network": "192.0.2.2/32",
"source-ipv4-network": "203.0.113.3/32",
"protocol":17
},
"udp": {
"source-port": {
"operator": "eq",
"port": 53
}
}
},
"actions": {
"forwarding": "drop"
}
}
]
}
}
]
}
}
Figure 4: JSON Example of ACL including (src_ip, src_port, dst_ip, dst_port, protocol) information conveyed by DOTS data channel
In this case, the link can be congested when it is under volume attack-time, so DOTS data channel can't be used to convey the drop-listed filtering rules as blocked traffic information [Interop]. On the other hand, DOTS signal channel [I-D.ietf-dots-signal-channel], the source-* clauses defined in [I-D.ietf-dots-signal-call-home] and filtering control [I-D.ietf-dots-signal-filter-control] can be used to communicate the policies to the orchestrator.
DOTS signal channel has capability to send (dst_ip, dst_port, protocol) information. The procedure to use DOTS Signal Channel in this case is as follows:
{
"ietf-dots-signal-channel:mitigation-scope": {
"scope": [
{
"target-prefix": [
"192.0.2.2/32"
],
"target-port-range": [
{
"lower-port": 80
},
{
"lower-port": 443
}
],
"target-protocol": [
6
],
"lifetime": 3600
},
{
"target-prefix": [
"192.0.2.2/32"
],
"target-port-range": [
{
"lower-port": 53
},
{
"lower-port": 123
}
],
"target-protocol": [
17
],
"lifetime": 3600
}
]
}
}
Figure 5: JSON Example of offload request including (dst_ip, dst_port, protocol) information conveyed by DOTS signal channel
[I-D.ietf-dots-signal-call-home] extends the DOTS signal channel to convey (dst_ip, dst_port, src_ip, src_port, protocol) information in a mitigation request. A mitigation request can convey src_ip information when the number of reflectors detected by a DMS is small. The procedure to use DOTS src-* clauses is as follows:
{
"ietf-dots-signal-channel:mitigation-scope": {
"scope": [
{
"target-prefix": [
"192.0.2.2/32"
],
"target-protocol": [
6
],
"ietf-dots-call-home:source-prefix": [
"203.0.113.2/32"
],
"ietf-dots-call-home:source-port-range" : [
{
"ietf-dots-call-home:lower-port": 53
},
{
"ietf-dots-call-home:lower-port": 123
}
],
"lifetime": 3600
},
{
"target-prefix": [
"192.0.2.2/32"
],
"target-protocol": [
6
],
"ietf-dots-call-home:source-prefix": [
"203.0.113.3/32"
],
"ietf-dots-call-home:source-port-range" : [
{
"ietf-dots-call-home:lower-port": 19
},
{
"ietf-dots-call-home:lower-port": 11211
}
],
"lifetime": 3600
}
]
}
}
Figure 6: JSON Example of offload request including (dst_ip, src_ip, src_port, protocol) information conveyed by DOTS signal channel
On the other hand, a mitigation request cannot convey src_ip information when number of reflector detected by DMS exceeds a certain number (cannot fit within one single request). The procedure to use the DOTS signal channel in the situation is as follows:
{
"ietf-dots-signal-channel:mitigation-scope": {
"scope": [
{
"target-prefix": [
"192.0.2.2/32"
],
"target-protocol": [
6
],
"ietf-dots-call-home:source-port-range" : [
{
"ietf-dots-call-home:lower-port": 53
},
{
"ietf-dots-call-home:lower-port": 123
},
{
"ietf-dots-call-home:lower-port": 19
},
{
"ietf-dots-call-home:lower-port": 11211
}
],
"lifetime": 3600
}
]
}
}
Figure 7: JSON Example of offload request including (dst_ip, src_port, protocol) information conveyed by DOTS signal channel
DOTS signal channel controlling filtering [I-D.ietf-dots-signal-filter-control] has capability to activate or deactivate ACL configured by Data Channel. Against reflection attack, DOTS client configures ACL including (dst_ip, src_port, protocol) information in peace time by Data Channel, and DOTS client activate the ACL in attack time by Signal Channel controlling filtering. Note that the src_port is well known port abused to carry out reflection attack by attacker. The procedure to use DOTS data channel and signal channel controlling filtering is as follows:
{
"ietf-dots-data-channel:acls": {
"acl": [
{
"name": "DMS_Offload_scenario_ACL",
"type": "ipv4-acl-type",
"activation-type": "activate-when-mitigating",
"aces": {
"ace": [
{
"name": "DMS_Offload_scenario_ACL_DNS_amp",
"matches": {
"ipv4": {
"destination-ipv4-network": "192.0.2.2/32",
"protocol":17
},
"udp": {
"source-port": {
"operator": "eq",
"port": 53
}
}
},
"actions": {
"forwarding": "drop"
}
},
{
"name": "DMS_Offload_scenario_ACL_NTP_amp",
"matches": {
"ipv4": {
"destination-ipv4-network": "192.0.2.2/32",
"protocol":17
},
"udp": {
"source-port": {
"operator": "eq",
"port": 123
}
}
},
"actions": {
"forwarding": "drop"
}
}
]
}
}
]
}
}
Figure 8: JSON Example of ACL including (dst_ip, src_port, protocol) information conveyed by DOTS data channel
{
"ietf-dots-signal-channel:mitigation-scope": {
"scope": [
{
"target-prefix": [
"192.0.2.2/32"
],
"target-protocol": [
17
],
"acl-list": [
{
"acl-name": "DMS_Offload_scenario_ACL_DNS_amp",
"activation-type": "immediate"
}
"lifetime": 3600
}
]
}
}
Figure 9: JSON Example of including acl name conveyed by DOTS signal channel
Against non-reflection attack, DOTS client configures ACL including (dst_ip, dst_port, protocol) information in peace time by Data Channel, and DOTS client activate the 'acl' in attack time by Signal Channel. Note that the dst_port is well known port abused to carry out non-reclection attack by attacker. The procedure to use DOTS data channel and signal channel controlling filtering is as follows:
{
"ietf-dots-data-channel:acls": {
"acl": [
{
"name": "DMS_Offload_scenario_ACL",
"type": "ipv4-acl-type",
"activation-type": "activate-when-mitigating",
"aces": {
"ace": [
{
"name": "DMS_Offload_scenario_HTTP_GET_Flooding",
"matches": {
"ipv4": {
"destination-ipv4-network": "192.0.2.2/32",
"protocol":6
},
"tcp": {
"destination-port": {
"operator": "eq",
"port": 80
}
}
},
"actions": {
"forwarding": "drop"
}
},
{
"name": "DMS_Offload_scenario_SYN_Flooding_FTP",
"matches": {
"ipv4": {
"destination-ipv4-network": "192.0.2.2/32",
"protocol":6
},
"tcp": {
"destination-port": {
"operator": "eq",
"port": 20
}
}
},
"actions": {
"forwarding": "drop"
}
}
]
}
}
]
}
}
Figure 10: JSON Example of ACL including (dst_ip, dst_port, protocol) information conveyed by DOTS data channel
{
"ietf-dots-signal-channel:mitigation-scope": {
"scope": [
{
"target-prefix": [
"192.0.2.2/32"
],
"target-protocol": [
6
],
"acl-list": [
{
"acl-name": "DMS_Offload_scenario_HTTP_GET_Flooding",
"activation-type": "immediate"
}
"lifetime": 3600
}
]
}
}
Figure 11: JSON Example of including ACL name conveyed by DOTS signal channel
Security considerations discussed in [I-D.ietf-dots-data-channel] and [I-D.ietf-dots-signal-channel] are to be taken into account.
This document does not require any action from IANA.
Thanks to Tirumaleswar Reddy, Shunsuke Homma, Pan Wei, and Xia Liang for the comments.
Thanks to Koichi Sakurada for demonstrating proof of concepts of this proposal .
| [I-D.ietf-dots-data-channel] | Boucadair, M. and R. K, "Distributed Denial-of-Service Open Threat Signaling (DOTS) Data Channel Specification", Internet-Draft draft-ietf-dots-data-channel-30, July 2019. |
| [I-D.ietf-dots-signal-call-home] | K, R., Boucadair, M. and J. Shallow, "Distributed Denial-of-Service Open Threat Signaling (DOTS) Signal Channel Call Home", Internet-Draft draft-ietf-dots-signal-call-home-03, July 2019. |
| [I-D.ietf-dots-signal-channel] | K, R., Boucadair, M., Patil, P., Mortensen, A. and N. Teague, "Distributed Denial-of-Service Open Threat Signaling (DOTS) Signal Channel Specification", Internet-Draft draft-ietf-dots-signal-channel-35, July 2019. |
| [I-D.ietf-dots-signal-filter-control] | Nishizuka, K., Boucadair, M., K, R. and T. Nagata, "Controlling Filtering Rules Using Distributed Denial-of-Service Open Threat Signaling (DOTS) Signal Channel", Internet-Draft draft-ietf-dots-signal-filter-control-01, May 2019. |
| [I-D.ietf-dots-use-cases] | Dobbins, R., Migault, D., Fouant, S., Moskowitz, R., Teague, N., Xia, L. and K. Nishizuka, "Use cases for DDoS Open Threat Signaling", Internet-Draft draft-ietf-dots-use-cases-18, July 2019. |
| [Interop] | Nishizuka, K., Shallow, J. and L. Xia , "DOTS Interop test report, IETF 103 Hackathon", November 2018. |
| [RFC4271] | Rekhter, Y., Li, T. and S. Hares, "A Border Gateway Protocol 4 (BGP-4)", RFC 4271, DOI 10.17487/RFC4271, January 2006. |
| [RFC5575] | Marques, P., Sheth, N., Raszuk, R., Greene, B., Mauch, J. and D. McPherson, "Dissemination of Flow Specification Rules", RFC 5575, DOI 10.17487/RFC5575, August 2009. |
| [RFC8612] | Mortensen, A., Reddy, T. and R. Moskowitz, "DDoS Open Threat Signaling (DOTS) Requirements", RFC 8612, DOI 10.17487/RFC8612, May 2019. |