Distributed Denial-of-Service Open
Threat Signaling (DOTS) Signal ChannelCisco Systems, Inc.Cessna Business Park, Varthur HobliSarjapur Marathalli Outer Ring RoadBangaloreKarnataka560103Indiatireddy@cisco.comOrangeRennes35000Francemohamed.boucadair@orange.comCisco Systems, Inc.170 West Tasman DriveSan JoseCalifornia95134USAdwing@cisco.comCisco Systems, Inc.praspati@cisco.comDOTSThis document specifies a mechanism that a DOTS client can use to
signal that a network is under a Distributed Denial-of-Service (DDoS)
attack to an upstream DOTS server so that appropriate mitigation actions
are undertaken (including, blackhole, drop, rate-limit, or add to watch
list) on the suspect traffic. The document specifies the DOTS signal
channel including Happy Eyeballs considerations. The specification of
the DOTS data channel is elaborated in a companion document.A distributed denial-of-service (DDoS) attack is an attempt to make
machines or network resources unavailable to their intended users. In
most cases, sufficient scale can be achieved by compromising enough
end-hosts and using those infected hosts to perpetrate and amplify the
attack. The victim in this attack can be an application server, a host,
a router, a firewall, or an entire network.In many cases, it may not be possible for an enterprise network
administrators to determine the causes of an attack, but instead just
realize that certain resources seem to be under attack. This document,
which adheres to the DOTS architecture , proposes that, in such
cases, the DOTS client just inform its DOTS server(s) that the
enterprise is under a potential attack and that the mitigator monitor
traffic to the enterprise to mitigate any possible attacks. This
cooperation between DOTS agents contributes to ensure a highly automated
network that is also robust, reliable and secure.Protocol requirements for DOTS signal channel are obtained from DOTS
requirements .This document satisfies all the use cases discussed in except the Third-party DOTS
notifications use case in Section 3.2.3 of which is an optional feature
and not a core use case. Third-party DOTS notifications are not part of
the DOTS requirements document. Moreover, the DOTS architecture does not
assess whether that use case may have an impact on the architecture
itself and/or the DOTS trust model.This is a companion document to the DOTS data channel specification
[I-D.reddy-dots-data-channel].The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in .(D)TLS: For brevity this term is used for statements that apply to
both Transport Layer Security and
Datagram Transport Layer Security .
Specific terms will be used for any statement that applies to either
protocol alone.The reader should be familiar with the terms defined in .Network applications have finite resources like CPU cycles, number of
processes or threads they can create and use, maximum number of
simultaneous connections it can handle, limited resources of the control
plane, etc. When processing network traffic, such applications are
supposed to use these resources to offer the intended task in the most
efficient fashion. However, an attacker may be able to prevent an
application from performing its intended task by causing the application
to exhaust the finite supply of a specific resource.TCP DDoS SYN-flood, for example, is a memory-exhaustion attack on the
victim and ACK-flood is a CPU exhaustion attack on the victim (). Attacks on the link are carried out by
sending enough traffic such that the link becomes excessively congested,
and legitimate traffic suffers high packet loss. Stateful firewalls can
also be attacked by sending traffic that causes the firewall to hold
excessive state and the firewall runs out of memory, and can no longer
instantiate the state required to pass legitimate flows. Other possible
DDoS attacks are discussed in .In each of the cases described above, the possible arrangements
between the DOTS client and DOTS server to mitigate the attack are
discussed in . An example
of network diagram showing a deployment of these elements is shown in
. Architectural relationships between involved
DOTS agents is explained in . In this example, the DOTS
server is operating on the access network.The DOTS server can also be running on the Internet, as depicted in
.In typical deployments, the DOTS client belongs to a different
administrative domain than the DOTS server. For example, the DOTS client
is a web server serving content owned and operated by an domain, while
the DOTS server is owned and operated by a different domain providing
DDoS mitigation services. That domain providing DDoS mitigation service
might, or might not, also provide Internet access service to the website
operator.The DOTS server may (not) be co-located with the DOTS mitigator. In
typical deployments, the DOTS server belongs to the same administrative
domain as the mitigator.The DOTS client can communicate directly with the DOTS server or
indirectly via a DOTS gateway.This document focuses on the DOTS signal channel.DOTS signaling can happen with DTLS
over UDP and TLS over TCP. A DOTS client
can use DNS to determine the IP address(es) of a DOTS server or a DOTS
client may be provided with the list of DOTS server IP addresses. The
DOTS client MUST know a DOTS server's domain name; hard-coding the
domain name of the DOTS server into software is NOT RECOMMENDED in case
the domain name is not valid or needs to change for legal or other
reasons. The DOTS client performs A and/or AAAA record lookup of the
domain name and the result will be a list of IP addresses, each of which
can be used to contact the DOTS server using UDP and TCP.If an IPv4 path to reach a DOTS server is found, but the DOTS
server's IPv6 path is not working, a dual-stack DOTS client can
experience a significant connection delay compared to an IPv4-only DOTS
client. The other problem is that if a middlebox between the DOTS client
and DOTS server is configured to block UDP, the DOTS client will fail to
establish a DTLS session with the DOTS server and will, then, have to
fall back to TLS over TCP incurring significant connection delays. discusses that DOTS client
and server will have to support both connectionless and
connection-oriented protocols.To overcome these connection setup problems, the DOTS client can try
connecting to the DOTS server using both IPv6 and IPv4, and try both
DTLS over UDP and TLS over TCP in a fashion similar to the Happy
Eyeballs mechanism . These connection
attempts are performed by the DOTS client when its initializes, and the
client uses that information for its subsequent alert to the DOTS
server. In order of preference (most preferred first), it is UDP over
IPv6, UDP over IPv4, TCP over IPv6, and finally TCP over IPv4, which
adheres to address preference order and
the DOTS preference that UDP be used over TCP (to avoid TCP's head of
line blocking).In reference to , the DOTS
client sends two TCP SYNs and two DTLS ClientHello messages at the same
time over IPv6 and IPv4. In this example, it is assumed that the IPv6
path is broken and UDP is dropped by a middle box but has little impact
to the DOTS client because there is no long delay before using IPv4 and
TCP. The IPv6 path and UDP over IPv6 and IPv4 is retried until the DOTS
client gives up.Constrained Application Protocol (CoAP) is used for DOTS signal channel (). COAP was designed according to the REST
architecture, and thus exhibits functionality similar to that of HTTP,
it is quite straightforward to map from CoAP to HTTP and from HTTP to
CoAP. CoAP has been defined to make use of both DTLS over UDP and TLS
over TCP. The advantages of COAP are: (1) Like HTTP, CoAP is based on
the successful REST model, (2) CoAP is designed to use minimal
resources, (3) CoAP integrates with JSON, CBOR or any other data
format, (4) asynchronous message exchanges, etc.A single DOTS signal channel between DOTS agents can be used to
exchange multiple DOTS signal messages. To reduce DOTS client and DOTS
server workload, DOTS client SHOULD re-use the (D)TLS session.JSON payloads are used to convey
signal channel specific payload messages that convey request
parameters and response information such as errors.TBD: Do we want to use CBOR [RFC7049] instead of JSON?The following APIs define the means to convey a DOTS signal from a
DOTS client to a DOTS server:are used to convey the DOTS signal
from a DOTS client to a DOTS server over the signal channel,
possibly traversing a DOTS gateway, indicating the DOTS client's
need for mitigation, as well as the scope of any requested
mitigation (). DOTS gateway act as a
CoAP-to-CoAP Proxy (explained in ).are used by the DOTS client to
withdraw the request for mitigation from the DOTS server ().are used by the DOTS client to
retrieve the DOTS signal(s) it had conveyed to the DOTS server
().are used by the DOTS client to convey
mitigation efficacy updates to the DOTS server ().Reliability is provided to the POST, DELETE, GET, and PUT requests
by marking them as Confirmable (CON) messages. As explained in Section
2.1 of , a Confirmable message is
retransmitted using a default timeout and exponential back-off between
retransmissions, until the DOTS server sends an Acknowledgement
message (ACK) with the same Message ID conveyed from the DOTS client.
Message transmission parameters are defined in Section 4.8 of . Reliability is provided to the responses by
marking them as Confirmable (CON) messages. The DOTS server can either
piggback the response in the acknowledgement message or if the DOTS
server is not able to respond immediately to a request carried in a
Confirmable message, it simply responds with an Empty Acknowledgement
message so that the DOTS client can stop retransmitting the request.
Empty Acknowledgement message is explained in Section 2.2 of . When the response is ready, the server sends
it in a new Confirmable message which then in turn needs to be
acknowledged by the DOTS client (see Sections 5.2.1 and Sections 5.2.2
in ).Implementation Note: A DOTS client that receives a response in a
CON message may want to clean up the message state right after sending
the ACK. If that ACK is lost and the DOTS server retransmits the CON,
the DOTS client may no longer have any state to which to correlate
this response, making the retransmission an unexpected message; the
DOTS client will send a Reset message so it does not receive any more
retransmissions. This behavior is normal and not an indication of an
error (see Section 5.3.2 in for more
details).When suffering an attack and desiring DoS/DDoS mitigation, a DOTS
signal is sent by the DOTS client to the DOTS server. A POST request
is used to convey a DOTS signal to the DOTS server (). The DOTS server can enable mitigation on
behalf of the DOTS client by communicating the DOTS client's request
to the mitigator and relaying any mitigator feedback to the
requesting DOTS client.The header fields are described below.Identifier of the policy represented
using an integer. This identifier MUST be unique for each policy
bound to the DOTS client, i.e. ,the policy-id needs to be unique
relative to the active policies with the DOTS server. This
identifier MUST be generated by the DOTS client. This document
does not make any assumption about how this identifier is
generated. This is a mandatory attribute.A list of IP addresses or prefixes
under attack. IP addresses and prefixes are separated by commas.
Prefixes are represented using CIDR notation . This is an optional attribute.A list of ports under attack. Ports
are separated by commas and port number range (using "-"). For
TCP, UDP, SCTP, or DCCP: the range of ports (e.g., 1024-65535).
This is an optional attribute.A list of protocols under attack.
Valid protocol values include tcp, udp, sctp, and dccp. Protocol
values are separated by commas. This is an optional
attribute.Name of the alias (see Section 3.1.1 in
[I-D.reddy-dots-data-channel). This is an optional
attribute.Lifetime of the mitigation request
policy in seconds. Upon the expiry of this lifetime, and if the
request is not refreshed, the mitigation request is removed. The
request can be refreshed by sending the same request again. The
default lifetime of the policy is 60 minutes -- this value was
chosen to be long enough so that refreshing is not typically a
burden on the DOTS client, while expiring the policy where the
client has unexpectedly quit in a timely manner. A lifetime of
zero indicates indefinite lifetime for the mitigation request.
The server MUST always indicate the actual lifetime in the
response. This is an optional attribute in the request.The relative order of two rules is determined by comparing their
respective policy identifiers. The rule with lower numeric policy
identifier value has higher precedence (and thus will match before)
than the rule with higher numeric policy identifier value.To avoid DOTS signal message fragmentation and the consequently
decreased probability of message delivery, DOTS agents MUST ensure
that the DTLS record MUST fit within a single datagram. If the Path
MTU is not known to the DOTS server, an IP MTU of 1280 bytes SHOULD
be assumed. The length of the URL MUST NOT exceed 256 bytes. If UDP
is used to convey the DOTS signal and the request size exceeds the
Path MTU then the DOTS client MUST split the DOTS signal into
separate messages, for example the list of addresses in the
'target-ip' field could be split into multiple lists and each list
conveyed in a new POST request.Implementation Note: DOTS choice of message size parameters works
well with IPv6 and with most of today's IPv4 paths. However, with
IPv4, it is harder to absolutely ensure that there is no IP
fragmentation. If IPv4 support on unusual networks is a
consideration and path MTU is unknown, implementations may want to
limit themselves to more conservative IPv4 datagram sizes such as
576 bytes, as per IP packets up to
576 bytes should never need to be fragmented, thus sending a maximum
of 500 bytes of DOTS signal over a UDP datagram will generally avoid
IP fragmentation. shows a POST request to signal
that ports 80, 8080, and 443 on the servers 2002:db8:6401::1 and
2002:db8:6401::2 are being attacked.The DOTS server indicates the result of processing the POST
request using CoAP response codes. CoAP 2xx codes are success, CoAP
4xx codes are some sort of invalid requests and 5xx codes are
returned if the DOTS server has erred or is incapable of performing
the mitigation. Response code 2.01 (Created) will be returned in the
response if the DOTS server has accepted the mitigation request and
will try to mitigate the attack. If the request is missing one or
more mandatory attributes, then 4.00 (Bad Request) will be returned
in the response or if the request contains invalid or unknown
parameters then 4.02 (Invalid query) will be returned in the
response. The CoAP response will include the JSON body received in
the request.A DELETE request is used to withdraw a DOTS signal from a DOTS
server ().If the DOTS server does not find the policy number conveyed in
the DELETE request in its policy state data, then it responds with a
4.04 (Not Found) error response code. The DOTS server successfully
acknowledges a DOTS client's request to withdraw the DOTS signal
using 2.02 (Deleted) response code, and ceases mitigation activity
as quickly as possible.A GET request is used to retrieve information and status of a
DOTS signal from a DOTS server (). If
the DOTS server does not find the policy number conveyed in the GET
request in its policy state data, then it responds with a 4.04 (Not
Found) error response code. shows the response of all the
active policies on the DOTS server.The various possible values of status field are explained
below:Attack mitigation is in
progress (e.g., changing the network path to re-route the
inbound traffic to DOTS mitigator).Attack is successfully
mitigated (e.g., attack traffic is dropped).Attack has stopped and the DOTS
client can withdraw the mitigation request.The observe option defined in
extends the CoAP core protocol with a mechanism for a CoAP client to
"observe" a resource on a CoAP server: the client retrieves a
representation of the resource and requests this representation be
updated by the server as long as the client is interested in the
resource. A DOTS client conveys the observe option set to 0 in the
GET request to receive unsolicited notifications of attack
mitigation status from the DOTS server. Unidirectional notifications
within the bidirectional signal channel allows unsolicited message
delivery, enabling asynchronous notifications between the agents. A
DOTS client that is no longer interested in receiving notifications
from the DOTS server can simply "forget" the observation. When the
DOTS server then sends the next notification, the DOTS client will
not recognize the token in the message and thus will return a Reset
message. This causes the DOTS server to remove the associated
entry.A DOTS client retrieves the information about a DOTS signal at
frequent intervals to determine the status of an attack. If the
DOTS server has been able to mitigate the attack and the attack
has stopped, the DOTS server indicates as such in the status, and
the DOTS client recalls the mitigation request.A DOTS client should react to the status of the attack from the
DOTS server and not the fact that it has recognized, using its own
means, that the attack has been mitigated. This ensures that the
DOTS client does not recall a mitigation request in a premature
fashion because it is possible that the DOTS client does not sense
the DDOS attack on its resources but the DOTS server could be
actively mitigating the attack and the attack is not completely
averted.While DDoS mitigation is active, a DOTS client MAY frequently
transmit DOTS mitigation efficacy updates to the relevant DOTS
server. An PUT request () is used to
convey the mitigation efficacy update to the DOTS server. The PUT
request MUST include all the header fields used in the POST request
to convey the DOTS signal (). If the DOTS
server does not find the policy number conveyed in the PUT request
in its policy state data, it responds with a 4.04 (Not Found) error
response code.The 'attack-status' field is a mandatory attribute. The various
possible values contained in the 'attack-status' field are explained
below:DOTS client determines that it is
still under attack.Attack is successfully mitigated
(e.g., attack traffic is dropped).This section defines the (D)TLS protocol profile of DOTS signal
channel over (D)TLS and DOTS data channel over TLS.There are known attacks on (D)TLS, such as machine-in-the-middle and
protocol downgrade. These are general attacks on (D)TLS and not specific
to DOTS over (D)TLS; please refer to the (D)TLS RFCs for discussion of
these security issues. DOTS agents MUST adhere to the (D)TLS
implementation recommendations and security considerations of except with respect to (D)TLS version. Since
encryption of DOTS using (D)TLS is virtually a green-field deployment
DOTS agents MUST implement only (D)TLS 1.2 or later.Implementations compliant with this profile MUST implement all of the
following items:DOTS client can use (D)TLS session resumption without server-side
state to resume session and convey
the DOTS signal.While the communication to the DOTS server is quiescent, the DOTS
client MAY probe the server to ensure it has maintained
cryptographic state. Such probes can also keep alive firewall or NAT
bindings. This probing reduces the frequency of needing a new
handshake when a DOTS signal needs to be conveyed to the DOTS
server. A (D)TLS heartbeat verifies the
DOTS server still has DTLS state by returning a DTLS message. If
the server has lost state, it returns a DTLS Alert. Upon receipt
of an unauthenticated DTLS Alert, the DTLS client validates the
Alert is within the replay window (Section 4.1.2.6 of ). It is difficult for the DTLS client
to validate the DTLS Alert was generated by the DTLS server in
response to a request or was generated by an on- or off-path
attacker. Thus, upon receipt of an in-window DTLS Alert, the
client SHOULD continue re-transmitting the DTLS packet (in the
event the Alert was spoofed), and at the same time it SHOULD
initiate DTLS session resumption.TLS runs over TCP, so a simple probe is a 0-length TCP packet
(a "window probe"). This verifies the TCP connection is still
working, which is also sufficient to prove the server has
retained TLS state, because if the server loses TLS state it
abandons the TCP connection. If the server has lost state, a TCP
RST is returned immediately.Raw public keys which reduce
the size of the ServerHello, and can be used by servers that
cannot obtain certificates (e.g., DOTS gateways on private
networks).Implementations compliant with this profile SHOULD implement all of
the following items to reduce the delay required to deliver a DOTS
signal:TLS False Start
which reduces round-trips by allowing the TLS second flight of
messages (ChangeCipherSpec) to also contain the DOTS signal.Cached Information Extension which avoids transmitting
the server's certificate and certificate chain if the client has
cached that information from a previous TLS handshake.TCP Fast Open can reduce the
number of round-trips to convey DOTS signal.(D)TLS based on client certificate can be used for mutual
authentication between DOTS agents. If a DOTS gateway is involved, DOTS
clients and DOTS gateway MUST perform mutual authentication; only
authorized DOTS clients are allowed to send DOTS signals to a DOTS
gateway. DOTS gateway and DOTS server MUST perform mutual
authentication; DOTS server only allows DOTS signals from authorized
DOTS gateway, creating a two-link chain of transitive authentication
between the DOTS client and the DOTS server.In the example depicted in ,
the DOTS gateway and DOTS clients within the 'example.com' domain
mutually authenticate with each other. After the DOTS gateway validates
the identity of a DOTS client, it communicates with the AAA server in
the 'example.com' domain to determine if the DOTS client is authorized
to request DDOS mitigation. If the DOTS client is not authorized, a 4.01
(Unauthorized) is returned in the response to the DOTS client. In this
example, the DOTS gateway only allows the application server and DDOS
detector to request DDOS mitigation, but does not permit the user of
type 'guest' to request DDOS mitigation.Also, DOTS gateway and DOTS server MUST perform mutual authentication
using certificates. A DOTS server will only allow a DOTS gateway with a
certificate for a particular domain to request mitigation for that
domain. In reference to , the DOTS server
only allows the DOTS gateway to request mitigation for 'example.com'
domain and not for other domains.TODO[TBD: DOTS WG will probably have to do something similar to
https://tools.ietf.org/html/rfc7519#section-10, create JSON DOTS claim
registry and register the JSON attributes defined in this
specification].Authenticated encryption MUST be used for data confidentiality and
message integrity. (D)TLS based on client certificate MUST be used for
mutual authentication. The interaction between the DOTS agents requires
Datagram Transport Layer Security (DTLS) and Transport Layer Security
(TLS) with a cipher suite offering confidentiality protection and the
guidance given in MUST be followed to
avoid attacks on (D)TLS.If TCP is used between DOTS agents, an attacker may be able to inject
RST packets, bogus application segments, etc., regardless of whether TLS
authentication is used. Because the application data is TLS protected,
this will not result in the application receiving bogus data, but it
will constitute a DoS on the connection. This attack can be countered by
using TCP-AO . If TCP-AO is used, then any
bogus packets injected by an attacker will be rejected by the TCP-AO
integrity check and therefore will never reach the TLS layer.Special care should be taken in order to ensure that the activation
of the proposed mechanism won't have an impact on the stability of the
network (including connectivity and services delivered over that
network).Involved functional elements in the cooperation system must establish
exchange instructions and notification over a secure and authenticated
channel. Adequate filters can be enforced to avoid that nodes outside a
trusted domain can inject request such as deleting filtering rules.
Nevertheless, attacks can be initiated from within the trusted domain if
an entity has been corrupted. Adequate means to monitor trusted nodes
should also be enabled.The following individuals have contributed to this document:Mike Geller Cisco Systems, Inc. 3250 Florida 33309 USA Email:
mgeller@cisco.comRobert Moskowitz HTT Consulting Oak Park, MI 42837 United States
Email: rgm@htt-consult.comThanks to Christian Jacquenet, Roland Dobbins, Andrew Mortensen,
Roman D. Danyliw, and Gilbert Clark for the discussion and comments.