[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Dead peer detection
Hi Ed,
The draft has expired, but I've attached a copy of it. I'd like to move the
draft forward (wherever that might be), but the focus in the WG lately has
been on IKEv2.
> I have wondered around the working groups site and could not find the
> draft-ietf-ipsec-dpd-00.txt any longer nor could I find any on going
> conversations on the subject. Was this draft allowed to expire
> without any
> further discussions, or was another draft started.
> I understand that some products do "dead peer detection" and was wondering
> if this draft was how it was to be done or if the use of lower
> re-key timer
This is the method that Cisco devices use.
> (say 600 seconds) in phase one would have the same effect, if one was to
> delete the phase 2 sa's if the phase one negotiations failed.
It depends on your implementation. If you always maintain a Phase 1 SA
("Continuous Channel Mode") when there are Phase 2 SAs, then doing as you
propose might be one solution. Keep in mind, however, that it won't scale
well if you have to frequently rekey. Essentially, you'd be using a Phase 1
negotiation to do the DPD; the difference is that DPD involves 2 notify
messages (a "hello" and an ACK). Using a Phase 1 for DPD requires at least
3 messages, plus a DH...Added to this is the concern that 600 seconds might
be too coarse an interval before detecting a black hole.
-g
> Thanks
> Ed Wilkinson
>
IPSec Working Group
G. Huang
S. Beaulieu
Internet Draft D. Rochefort
Document: draft-ietf-ipsec-dpd-01.txt Cisco Systems, Inc.
Expires: January 2002 August 2001
A Traffic-Based Method of Detecting Dead IKE Peers
Status of this Memo
This document is an Internet-Draft and is in full conformance
with all provisions of Section 10 of RFC2026.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
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."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
Abstract
This draft describes a method of detecting a dead IKE peer. The
method, called Dead Peer Detection (DPD) uses IPSec traffic patterns
to limit the number of IKE messages sent. DPD, like other keepalive
mechanisms, is often necessary to perform IKE peer failover, or to
reclaim lost resources.
Table of Contents
Status of this Memo................................................1
Abstract...........................................................1
1. Introduction....................................................2
2. Conventions used in this document...............................3
3. Document Roadmap................................................3
4. Keepalives vs. Heartbeats.......................................3
4.1 Keepalives:..................................................3
4.2 Heartbeats:..................................................4
Huang, Beaulieu, Rochefort Expires January 2002 1
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
5. DPD Protocol....................................................5
5.1 DPD Vendor ID................................................6
5.2 Message Exchanges............................................6
5.3 NOTIFY(R-U-THERE/R-U-THERE-ACK) Message Format...............7
5.4 Impetus for DPD Exchange.....................................8
5.5 Implementation Suggestion....................................8
5.6 Comparisons..................................................8
6. Resistance to Replay Attack and False Proof of Liveliness.......9
6.1 Sequence Number in DPD Messages..............................9
6.2 Selection and Maintenance of Sequence Numbers................9
7. Acknowledgements................................................9
8. References......................................................9
9. Editors' Address................................................9
1. Introduction
When two peers communicate with IKE/IPSec, the situation may arise
in which connectivity between the two goes down unexpectedly. This
situation can arise because of routing problems, one host rebooting,
etc., and in such cases, there is often no way for IKE and IPSec to
identify the loss of peer connectivity. As such, the SAs can remain
until their lifetimes naturally expire, resulting in a "black hole"
situation where packets are tunneled to oblivion. It is often
desirable to recognize black holes as soon as possible so that an
entity can failover to a different peer quickly. Likewise, it is
sometimes necessary to detect black holes to recover lost resources.
This problem of detecting a dead IKE peer has been addressed by
proposals that require sending periodic HELLO/ACK messages to prove
liveliness. These schemes tend to be unidirectional (a HELLO only)
or bidirectional (a HELLO/ACK pair). For the purpose of this draft,
the term "heartbeat" will refer to a unidirectional message to prove
liveliness. Likewise, the term "keepalive" will refer to a
bidirectional message.
The problem with current heartbeat and keepalive proposals is their
reliance upon their messages to be sent at regular intervals. In
the implementation, this translates into managing some timer to
service these message intervals. Similarly, because rapid detection
of the dead peer is often desired, these messages must be sent with
some frequency, again translating into considerable overhead for
message processing. In implementations and installations where
managing large numbers of simultaneous IKE sessions is of concern,
these regular heartbeats/keepalives prove to be infeasible.
To this end, this draft proposes an approach to detect peer
liveliness without needing to send messages at regular intervals.
This scheme is called Dead Peer Detection (DPD).
Huang, Beaulieu, Rochefort Expires January 2002 2
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
2. Conventions used in this document
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 RFC-2119 [1].
3. Document Roadmap
As mentioned above, there are already proposed solutions to the
problem of detecting dead peers. Section 4 examines a keepalives-
based approach as well as a heartbeats-based approach. A brief
comparison is also made between schemes based on the IPSec SA and
schemes based on the IKE SA. DPD is an IKE SA-based scheme.
Section 5 presents the DPD proposal fully, highlighting differences
between DPD and the schemes presented in Section 4 and emphasizing
scalability issues. Section 6 examines security issues surrounding
replayed messages and false liveliness, drawing upon work already
presented in the Heartbeats draft [2].
4. Keepalives vs. Heartbeats
4.1 Keepalives:
Consider a keepalives scheme in which peer A and peer B require
regular acknowledgements of each other's liveliness. The messages
are exchanged by means of an authenticated notify payload. The two
peers must agree upon the interval at which keepalives are sent,
meaning that some negotiation is required during Phase 1. For any
prompt failover to be possible, the keepalives must also be sent at
rather frequent intervals -- around 10 seconds or so. In this
hypothetical keepalives scenario, peers A and B agree to exchange
keepalives every 10 seconds. Essentially, every 10 seconds, one
peer must send a HELLO to the other. This HELLO serves as proof of
liveliness for the sending entity. In turn, the other peer must
acknowledge each keepalive HELLO. If the 10 seconds elapse, and one
side has not received a HELLO, it will send the HELLO message
itself, using the peer's ACK as proof of liveliness. Receipt of
either a HELLO or ACK causes an entity's keepalive timer to reset.
Failure to receive an ACK in a certain period of time signals an
error. A clarification is presented below:
Scenario 1:
Peer A's 10-second timer elapses first, and it sends a HELLO to B.
B responds with an ACK.
Peer A: Peer B:
10 second timer fires; ------>
wants to know that B is alive;
sends HELLO.
Receives HELLO; acknowledges
A's liveliness;
<------ resets keepalive timer, sends
ACK.
Receives ACK as proof of
Huang, Beaulieu, Rochefort Expires January 2002 3
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
B's liveliness; resets timer.
Scenario 2:
Peer A's 10-second timer elapses first, and it sends a HELLO to B.
B fails to respond. A can retransmit, in case its initial HELLO is
lost. This situation describes how peer A detects its peer is dead.
Peer A: Peer B (dead):
10 second timer fires; ------X
wants to know that B is
alive; sends HELLO.
Retransmission timer ------X
expires; initial message
could have been lost in
transit; A increments
error counter and
sends another HELLO.
. . .
After some number of errors, A assumes B is dead; deletes SAs and
possibly initiates failover.
An advantage of this scheme is that the party interested in the
other peer's liveliness begins the message exchange. In Scenario 1,
peer A is interested in peer B's liveliness, and peer A consequently
sends the HELLO. It is conceivable in such a scheme that peer B
would never be interested in peer A's liveliness. In such a case,
the onus would always lie on peer A to initiate the exchange.
4.2 Heartbeats:
By contrast, consider a proof-of-liveliness scheme involving
unidirectional (unacknowledged) messages. An entity interested in
its peer's liveliness would rely on the peer itself to send periodic
messages demonstrating liveliness. In such a scheme, the message
exchange might look like this:
Scenario 3:
Peer A and Peer B are interested in each other's liveliness. Each
peer depends on the other to send periodic HELLOs.
Peer A: Peer B:
10 second timer fires; ------>
sends HELLO. Timer also
signals expectation of
B's HELLO.
Receives HELLO as proof of A's
liveliness.
<------ 10 second timer fires; sends
Huang, Beaulieu, Rochefort Expires January 2002 4
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
HELLO.
Receives HELLO as proof
of B's liveliness.
Scenario 4:
Peer A fails to receive HELLO from B and marks the peer dead. This
is how an entity detects its peer is dead.
Peer A: Peer B (dead):
10 second timer fires; ------X
sends HELLO. Timer also
signals expectation of
B's HELLO.
. . .
Some time passes and A assumes B is dead.
The disadvantage of this scheme is the reliance upon the peer to
demonstrate liveliness. To this end, peer B might never be
interested in peer A's liveliness. Nonetheless, if A is interested
B's liveliness, B must be aware of this, and maintain the necessary
state information to send periodic HELLOs to A. The disadvantage of
such a scheme becomes clear in the remote-access scenario. Consider
a VPN aggregator that terminates a large number of sessions (on the
order of 50,000 peers or so). Each peer requires fairly rapid
failover, therefore requiring the aggregator to send HELLO packets
every 10 seconds or so. Such a scheme simply lacks scalability, as
the aggregator must send 50,000 messages every few seconds.
In both of these schemes (keepalives and heartbeats), some
negotiation of message interval must occur, so that each entity can
know how often its peer expects a HELLO. This immediately adds a
degree of complexity. Similarly, the need to send periodic messages
(regardless of other IPSec/IKE activity), also increases
computational overhead to the system.
5. DPD Protocol
DPD addresses the shortcomings of IKE keepalives- and heartbeats-
schemes by introducing a more reasonable logic governing message
exchange. Essentially, keepalives and heartbeats mandate exchange
of HELLOs at regular intervals. By contrast, with DPD, each peer’s
DPD state is largely independent of the other’s. A peer is free to
request proof of liveliness when it needs it -- not at mandated
intervals. This asynchronous property of DPD exchanges allows fewer
messages to be sent, and this is how DPD achieves greater
scalability.
As an elaboration, consider two DPD peers A and B. If there is
ongoing IPSec traffic between the two, there is little need for
proof of liveliness. The IPSec traffic itself serves as the proof
of liveliness. If, on the other hand, a period of time lapses
Huang, Beaulieu, Rochefort Expires January 2002 5
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
during which no packet-exchange occurs, the liveliness of each peer
is questionable. Knowledge of the peer’s liveliness, however, is
only urgently necessary if there is traffic to be sent. For
example, if peer A has some IPSec packets to send after the period
of idleness, it will need to know if peer B is still alive. At this
point, peer A can initiate the DPD exchange.
To this end, each peer may have different requirements for detecting
proof of liveliness. Peer A, for example, may require rapid
failover, whereas peer B's requirements for resource cleanup are
less urgent. In DPD, each peer can define its own "worry metric" –
an interval that defines the urgency of the DPD exchange.
Continuing the example, peer A might define its DPD interval to be
10 seconds. Then, if peer A sends outbound IPSec traffic, but fails
to receive any inbound traffic for 10 seconds, it can initiate a DPD
exchange.
Peer B, on the other hand, defines its less urgent DPD interval to
be 5 minutes. If the IPSec session is idle for 5 minutes, peer B
can initiate a DPD exchange the next time it sends IPSec packets.
It is important to note that the decision about when to initiate a
DPD exchange is implementation specific. An implementation might
even define the DPD messages to be at regular intervals following
idle periods. See section 5.5 for more implementation suggestions.
5.1 DPD Vendor ID
To demonstrate DPD capability, an entity must send the DPD vendor
ID. Both peers of an IKE session MUST send the DPD vendor ID before
DPD exchanges can begin. The format of the DPD Vendor ID is:
1
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! !M!M!
! HASHED_VENDOR_ID !J!N!
! !R!R!
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
where HASHED_VENDOR_ID = {0xAF, 0xCA, 0xD7, 0x13, 0x68, 0xA1, 0xF1,
0xC9, 0x6B, 0x86, 0x96, 0xFC, 0x77, 0x57}, and MJR and MNR
correspond to the current major and minor version of this protocol
(1 and 0 respectively). An IKE peer MUST send the Vendor ID if it
wishes to take part in DPD exchanges.
5.2 Message Exchanges
The DPD exchange is a bidirectional (HELLO/ACK) Notify message. The
exchange is defined as:
Sender Responder
-------- -----------
HDR*, NOTIFY(R-U-THERE), HASH ------>
Huang, Beaulieu, Rochefort Expires January 2002 6
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
<------ HDR*, NOTIFY(R-U-THERE-
ACK), HASH
The R-U-THERE message corresponds to a "HELLO" and the R-U-THERE-ACK
corresponds to an "ACK." Both messages are simply ISAKMP Notify
payloads, and as such, this draft defines these two new ISAKMP
Notify message types:
Notify Message Value
R-U-THERE 36136
R-U-THERE-ACK 36137
An entity that has sent the DPD Vendor ID MUST respond to an R-U-
THERE query. Furthermore, an entity MUST reject unencrypted R-U-
THERE and R-U-THERE-ACK messages.
5.3 NOTIFY(R-U-THERE/R-U-THERE-ACK) Message Format
When sent, the R-U-THERE message MUST take the following form:
1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! Next Payload ! RESERVED ! Payload Length !
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! Domain of Interpretation (DOI) !
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! Protocol-ID ! SPI Size = 16 ! Notify Message Type !
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! !
~ Security Parameter Index (SPI) ~
! !
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! !
~ Notification Data ~
! !
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
As this message is an ISAKMP NOTIFY, the Next Payload, RESERVED, and
Payload Length fields should be set accordingly. The remaining
fields are set as:
- Domain of Interpretation (4 octets) - SHOULD be set to IPSEC-DOI.
- Protocol ID (1 octet) – MUST be set to the protocol ID for ISAKMP.
- SPI Size (2 octets) - SHOULD be set to sixteen (16), the length of
two octet-sized ISAKMP cookies.
- Notify Message Type (2 octets) - MUST be set to R-U-THERE
- Security Parameter Index (16 octets) - SHOULD be set to the
cookies of the Initiator and Responder of the IKE SA (in that
Huang, Beaulieu, Rochefort Expires January 2002 7
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
order)
- Notification Data (4 octets) - MUST be set to the sequence number
corresponding to this message
The format of the R-U-THERE-ACK message is the same, with the
exception that the Notify Message Type MUST be set to R-U-THERE-ACK.
Again, the Notification Data MUST be sent to the sequence number
corresponding to the received R-U-THERE message.
5.4 Impetus for DPD Exchange
Again, rather than relying on some negotiated time interval to force
the exchange of messages, DPD does not mandate the exchange of R-U-
THERE messages at any time. Instead, an IKE peer SHOULD send an R-
U-THERE query to its peer only if it is interested in the liveliness
of this peer. To this end, if traffic is regularly exchanged
between two peers, either peer SHOULD use this traffic as proof of
liveliness, and both peers SHOULD NOT initiate a DPD exchange.
5.5 Implementation Suggestion
Since the liveliness of a peer is only questionable when no traffic
is exchanged, a viable implementation might begin by monitoring
idleness. Along these lines, a peer's liveliness is only important
when there is outbound traffic to be sent. To this end, an
implementation can initiate a DPD exchange (i.e., send an R-U-THERE
message) when there has been some period of idleness, followed by
the desire to send outbound traffic. Likewise, an entity can
initiate a DPD exchange if it has sent outbound IPSec traffic, but
not received any inbound IPSec packets in response. A complete DPD
exchange (i.e., transmission of R-U-THERE and receipt of
corresponding R-U-THERE-ACK) will serve as proof of liveliness until
the next idle period.
Again, since DPD does not mandate any interval, this "idle period"
(or "worry metric") is left as an implementation decision. It is
not a negotiated value.
5.6 Comparisons
The performance benefit that DPD offers over traditional keepalives-
and heartbeats-schemes comes from the fact that regular messages do
not need to be sent. Thus, the need to maintain timer state is
mitigated, and the number of IKE messages to be processed is
reduced. As a consequence, the scalability of DPD is much better
than keepalives and heartbeats.
DPD maintains the HELLO/ACK model presented by keepalives, as it
follows that an exchange is initiated only by an entity interested
in the liveliness of its peer.
Huang, Beaulieu, Rochefort Expires January 2002 8
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
6. Resistance to Replay Attack and False Proof of Liveliness
6.1 Sequence Number in DPD Messages
To guard against message replay attacks and false proof of
liveliness, a 32-bit sequence number MUST be presented with each R-
U-THERE message. A responder to an R-U-THERE message MUST send an
R-U-THERE-ACK with the same sequence number. Upon receipt of the R-
U-THERE-ACK message, the initial sender SHOULD check the validity of
the sequence number. The initial sender SHOULD reject the R-U-
THERE-ACK if the sequence number fails to match the one sent with
the R-U-THERE message.
Additionally, both the receiver of the R-U-THERE and the R-U-THERE-
ACK message SHOULD check the validity of the Initiator and Responder
cookies presented in the SPI field of the payload.
6.2 Selection and Maintenance of Sequence Numbers
As both DPD peers can initiate a DPD exchange (i.e., both peers can
send R-U-THERE messages), each peer MUST maintain its own sequence
number for R-U-THERE messages. The first R-U-THERE message sent in
a session MUST be a randomly chosen number. To prevent rolling past
overflowing the 32-bit boundary, the high-bit of the sequence number
initially SHOULD be set to zero. Subsequent R-U-THERE messages MUST
increment the sequence number by one. Sequence numbers MAY reset at
the expiry of the IKE SA, moving to a newly chosen random number.
Each entity SHOULD also maintain its peer's R-U-THERE sequence
number, and an entity SHOULD reject the R-U-THERE message if it
fails to match the expected sequence number.
Implementations MAY maintain a window of acceptable sequence
numbers, but this specification makes no assumptions about how this
is done. Again, it is an implementation specific detail.
7. Acknowledgements
The authors of this draft would like to thank Andrew Krywaniuk and
Tero Kivinen for the idea of using a sequence number in keepalive
exchanges to prevent replay attacks. They presented this idea in
[2]. Likewise, many engineers at Cisco contributed to this
proposal, including Scott Fanning, Steve Goldhaber, Natalie Timms,
Jan Vilhuber, and Victor Volpe.
8. References
1 RFC 2119 Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997
2 Krywaniuk, A., Kivinen, T., "Using Isakmp Heartbeats for
Dead Peer Detection," <draft-ietf-ipsec-heartbeats-01.txt>
(work in progress), July 2000.
9. Editors' Address
Huang, Beaulieu, Rochefort Expires January 2002 9
A Traffic-Based Method of Detecting Dead IKE Peers August 2001
Geoffrey Huang
Cisco Systems, Inc.
170 West Tasman Drive
San Jose, CA 95134
Phone: (408) 525-5354
Email: ghuang@cisco.com
Stephane Beaulieu
Cisco Systems, Inc.
2000 Innovation Drive
Kanata, ON
Canada, K2K 3E8
Phone: (613) 271-3678
Email: stephane@cisco.com
Dany Rochefort
Cisco Systems, Inc.
124 Grove Street, Suite 205
Franklin, MA 02038
Phone: (508) 553-6136
Email: danyr@cisco.com
The IPsec working group can be contacted through the chairs:
Barbara Fraser
byfraser@cisco.com
Cisco Systems, Inc.
Ted T'so
tytso@mit.edu
Massachusetts Institute of Technology
Huang, Beaulieu, Rochefort Expires January 2002 10