Network Working Group P. Srisuresh
Request for Comments: 2888 Campio Communications
Category: Informational August 2000
Secure Remote Access with L2TP
Status of this Memo
This memo provides information for the Internet community. It does
not specify an Internet standard of any kind. Distribution of this
memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2000). All Rights Reserved.
Abstract
L2TP protocol is a virtual extension of PPP across IP network
infrastrUCture. L2TP makes possible for an access concentrator (LAC)
to be near remote clients, while allowing PPP termination server
(LNS) to be located in enterprise premises. L2TP allows an enterprise
to retain control of RADIUS data base, which is used to control
Authentication, Authorization and Accountability (AAA) of dial-in
users. The objective of this document is to extend security
characteristics of IPsec to remote access users, as they dial-in
through the Internet. This is accomplished without creating new
protocols and using the existing practices of Remote Access and
IPsec. Specifically, the document proposes three new RADIUS
parameters for use by the LNS node, acting as Secure Remote Access
Server (SRAS) to mandate network level security between remote
clients and the enterprise. The document also discusses limitations
of the approach.
1. Introduction and Overview
Now-a-days, it is common practice for employees to dial-in to their
enterprise over the PSTN (Public Switched Telephone Network) and
perform day-to-day operations just as they would if they were in
corporate premises. This includes people who dial-in from their home
and road warriors, who cannot be at the corporate premises. As the
Internet has become ubiquitous, it is appealing to dial-in through
the Internet to save on phone charges and save the dedicated voice
lines from being clogged with data traffic.
The document suggests an approach by which remote access over the
Internet could become a reality. The approach is founded on the
well-known techniques and protocols already in place. Remote Access
extensions based on L2TP, when combined with the security offered by
IPSec can make remote access over the Internet a reality. The
approach does not require inventing new protocol(s).
The trust model of remote access discussed in this document is viewed
principally from the perspective of an enterprise into which remote
access clients dial-in. A remote access client may or may not want to
enforce end-to-end IPsec from his/her end to the enterprise.
However, it is in the interest of the enterprise to mandate security
of every packet that it accepts from the Internet into the
enterprise. Independently, remote users may also pursue end-to-end
IPsec, if they choose to do so. That would be in addition to the
security requirement imposed by the enterprise edge device.
Section 2 has reference to the terminology used throughout the
document. Also mentioned are the limited scope in which some of these
terms may be used in this document. Section 3 has a brief description
of what constitutes remote access. Section 4 describes what
constitutes network security from an enterprise perspective. Section
5 describes the model of secure remote access as a viable solution to
enterprises. The solution presented in section 5 has some
limitations. These limitations are listed in section 6. Section 7 is
devoted to describing new RADIUS attributes that may be configured to
turn a NAS device into Secure Remote Access Server.
2. Terminology and scope
Definition of terms used in this document may be found in one of (a)
L2TP Protocol document [Ref 1], (b) IP security Architecture document
[Ref 5], or (c) Internet Key Exchange (IKE) document [Ref 8].
Note, the terms Network Access Server (NAS) and Remote Access
Server(RAS) are used interchangeably throughout the document. While
PPP may be used to carry a variety of network layer packets, the
focus of this document is limited to carrying IP datagrams only.
"Secure Remote Access Server" (SRAS) defined in this document refers
to a NAS that supports tunnel-mode IPsec with its remote clients.
Specifically, LNS is the NAS that is referred. Further, involuntary
tunneling is assumed for L2TP tunnel setup, in that remote clients
initiating PPP session and the LAC that tunnels the PPP sessions are
presumed to be distinct physical entities.
Lastly, there are a variety of transport mediums by which to tunnel
PPP packets between a LAC and LNS. Examples include Frame Relay or
ATM cloud and IP network infrastructure. For simplicity, the document
assumes a public IP infrastructure as the medium to transport PPP
packets between LAC and LNS. Security of IP packets (embedded within
PPP) in a trusted private transport medium is less of a concern for
the purposes of this document.
3. Remote Access operation
Remote access is more than mere authentication of remote clients by a
Network Access Server(NAS). Authentication, Authorization, Accounting
and routing are integral to remote access. A client must first pass
the authentication test before being granted link access to the
network. Network level services (such as IP) are granted based on the
authorization characteristics specified for the user in RADIUS.
Network Access Servers use RADIUS to scale for large numbers of users
supported. NAS also monitors the link status of the remote access
clients.
There are a variety of techniques by which remote access users are
connected to their enterprise and the Internet. At a link level, the
access techniques include ISDN digital lines, analog plain-old-
telephone-service lines, xDSL lines, cable and wireless to name a
few. PPP is the most common Layer-2 (L2)protocol used for carrying
network layer packets over these remote access links. PPP may be used
to carry a variety of network layer datagrams including IP, IPX and
AppleTalk. The focus of this document is however limited to IP
datagrams only.
L2TP is a logical extension of PPP over an IP infrastructure. While a
LAC provides termination of Layer 2 links, LNS provides the logical
termination of PPP. As a result, LNS becomes the focal point for (a)
performing the AAA operations for the remote users, (b) assigning IP
address and monitoring the logical link status (i.e., the status of
LAC-to-LNS tunnel and the link between remote user and LAC), and (c)
maintaining host-route to remote user network and providing routing
infrastructure into the enterprise.
L2TP uses control messages to establish, terminate and monitor the
status of the logical PPP sessions (from remote user to LNS). These
are independent of the data messages. L2TP data messages contain an
L2TP header, followed by PPP packets. The L2TP header identifies the
PPP session (amongst other things) to which the PPP packet belongs.
The IP packets exchanged from/to the remote user are carried within
the PPP packets. The L2TP data messages, carrying end-to-end IP
packets in an IP transport medium may be described as follows. The
exact details of L2TP protocol may be found in [Ref 1].
+----------------------+
IP Header
(LAC <->LNS)
+----------------------+
UDP Header
+----------------------+
L2TP Header
(incl. PPP Sess-ID)
+----------------------+
PPP Header
(Remote User<->LNS)
+----------------------+
End-to-end IP packet
(to/from Remote User)
+----------------------+
4. Requirements of an enterprise Security Gateway
Today's enterprises are aware of the various benefits of connecting
to the Internet. Internet is a vast source of Information and a means
to disseminate information and make available certain resources to
the external world. However, enterprises are also aware that security
breaches (by being connected to the Internet) can severely jeopardize
internal network.
As a result, most enterprises restrict access to a pre-defined set of
resources for external users. Typically, enterprises employ a
firewall to restrict access to internal resources and place
externally accessible servers in the DeMilitarized Zone (DMZ), in
front of the firewall, as described below in Figure 1.
----------------
( )
( )
( Internet )
( )
(_______________ )
WAN
.........\....
+-----------------+
Enterprise Router
+-----------------+
DMZ - Network
---------------------------------
+--+ +--+ +----------+
__ __ Firewall
/____\ /____\ +----------+
DMZ-Name DMZ-Web ...
Server Server
------------------
( )
( Internal Network )
( (private to the )
( enterprise) )
(_________________ )
Figure 1: Security model of an Enterprise using Firewall
Network Access Servers used to allow direct dial-in access (through
the PSTN) to employees are placed within the private enterprise
network so as to avoid access restrictions imposed by a firewall.
With the above model, private resources of an enterprise are
restricted for access from the Internet. Firewall may be configured
to occasionally permit access to a certain resource or service but is
not recommended on an operational basis as that could constitute a
security threat to the enterprise. It is of interest to note that
even when the firewall is configured to permit access to internal
resources from pre-defined external node(s), many internal servers,
such as NFS, enforce address based authentication and do not co-
operate when the IP address of the external node is not in corporate
IP address domain. In other Words, with the above security model, it
becomes very difficult to allow employees to access corporate
resources, via the Internet, even if you are willing to forego
security over the Internet.
With the advent of IPsec, it is possible to secure corporate data
across the Internet by employing a Security Gateway within the
enterprise. Firewall may be configured to allow IKE and IPsec packets
directed to a specific Security Gateway behind the firewall. It then
becomes the responsibility of the Security Gateway to employ the
right access list for external connections seeking entry into the
enterprise. Essentially, the access control functionality for IPsec
secure packets would be shifted to the Security Gateway (while the
access control for clear packets is retained with the firewall). The
following figure illustrates the model where a combination of
Firewall and Security Gateway control access to internal resources.
------------
( )
( )
( Internet )
( )
(___________ )
WAN
.........\....
+-----------------+
Enterprise Router
+-----------------+
DMZ - Network
------------------------------------------------------------
+--+ +--+ +----------+
__ __ Firewall
/____\ /____\ +----------+
DMZ-Name DMZ-Web ...
Server Server etc. LAN
------------------------------------
+----------+ +------------------+
LNS Security Gateway
Server (SGW)
+----------+ +------------------+
------------------
( )
( Internal Network )
( (Private to the )
( enterprise) )
(_________________ )
Figure 2: Security Model based on Firewall and Security Gateway
In order to allow employee dial-in over the Internet, an LNS may be
placed behind a firewall, and the firewall may be configured to allow
UDP access to the LNS from the Internet. Note, it may not be possible
to know all the IP addresses of the LACs located on the Internet at
configuration time. Hence, the need to allow UDP access from any node
on the Internet. The LNS may be configured to process only the L2TP
packets and drop any UDP packets that are not L2TP.
Such a configuration allows remote access over the Internet. However,
the above setup is prone to a variety of security attacks over the
Internet. It is easy for someone on the Internet to steal a remote
access session and gain access to precious resources of the
enterprise. Hence it is important that all packets are preserved with
IPsec to a security Gateway (SGW) behind the LNS, so the Security
Gateway will not allow IP packets into corporate network unless it
can authenticate the same.
The trust model of secure remote access assumes that the enterprise
and the end user are trusted domains. Everything in between is not
trusted. Any examination of the end-to-end packets by the nodes
enroute would violate this trust model. From this perspective, even
the LAC node enroute must not be trusted with the end-to-end IP
packets. Hence, location and operation of LAC is not relevant for the
discussion on security. On the other hand, location and operation of
LNS and the Security Gateway (SGW) are precisely the basis for
discussion.
Having security processing done on an independent Security gateway
has the following shortcomings.
1. Given the trust model for remote access, the SGW must be
configured with a set of security profiles, access control lists
and IKE authentication parameters for each user. This mandates an
independent provisioning of security parameters on a per-user
basis. This may not be able to take advantage of the user-centric
provisioning on RADIUS, used by the LNS node.
2. Unlike the LNS, SGW may not be in the routing path of remote
access packets. I.e., there is no guarantee that the egress IP
packets will go through the chain of SGW and LNS before they are
delivered to remote user. As a result, packets may be subject to
IPSec in one direction, but not in the other. This can be a
significant threat to the remote access trust model.
3. Lastly, the SGW node does not have a way to know when a remote
user node(s) simply died or the LAC-LNS tunnel failed. Being
unable to delete the SAs for users that no longer exist could
drain the resources of the SGW. Further, the LNS cannot even
communicate the user going away to the SGW because, the SGW
maintains its peer nodes based on IKE user ID, which could be
different the user IDs employed by the LNS node.
5. Secure Remote Access
Combining the functions of IPsec Security Gateway and LNS into a
single system promises to offer a viable solution for secure remote
access. By doing this, remote access clients will use a single node
as both (a) PPP termination point providing NAS service, and (b) the
Security gateway node into the enterprise. We will refer this node as
"Secure Remote Access Server" (SRAS).
The SRAS can benefit greatly from the confluence of PPP session and
IPsec tunnel end points. PPP session monitoring capability of L2TP
directly translates to being able to monitor IPsec tunnels. Radius
based user authorization ability could be used to configure the
security characteristics for IPsec tunnel. This includes setting
access control filters and security preferences specific to each
user. This may also be extended to configuring IKE authentication and
other negotiation parameters, when automated key exchange is
solicited. Security attributes that may be defined in Radius are
discussed in detail in section 7. Needless to say, the centralized
provisioning capability and scalability of Radius helps in the
configuration of IPsec.
As for remote access, the benefit is one of IPsec security as
befitting the trust model solicited by enterprises for the end-to-end
IP packets traversing the Internet. You may use simply AH where there
is no fear of external eaves-dropping, but you simply need to
authenticate packet data, including the source of packet. You may use
ESP (including ESP-authentication), where there is no trust of the
network and you do not want to permit eaves-dropping on corporate
activities.
Operation of SRAS requires that the firewall be configured to permit
UDP traffic into the SRAS node. The SRAS node in turn will process
just the L2TP packets and drop the rest. Further, the SRAS will
require all IP packets embedded within PPP to be one of AH and ESP
packets, directed to itself. In addition, the SRAS will also permit
IKE UDP packets (with source and destination ports sets to 500)
directed to itself in order to perform IKE negotiation and generate
IPsec keys dynamically. All other IP packets embedded within PPP will
be dropped. This enforces the security policy for the enterprise by
permitting only the secure remote access packets into the enterprise.
When a PPP session is dropped, the IPsec and ISAKMP SAs associated
with the remote access user are dropped from the SRAS. All the
shortcomings listed in the previous section with LNS and SGW on two
systems disappear withe Secure Remote Access Server. Figure 3 below
is a typical description of an enterprise supporting remote access
users using SRAS system.
------------
Remote Access +-------------+ ( )
+--+______ Link Local Access ( )
__ /___________ Concentrator----( Internet )
/____\ (LAC) ( )
RA-Host +-------------+ (____________)
WAN
.........\....
+-----------------+
Enterprise Router
+-----------------+
DMZ - Network
------------------------------------------
+--+ +--+ +----------+
__ __ Firewall
/____\ /____\ +----------+
DMZ-Name DMZ-Web ...
Server Server etc. LAN
------------------------------------
+---------------+
Secure Remote
Access Server
(SRAS)
+---------------+
---------------------
( )
+--+ ( Internal Network )
__------( (Private to the )
/____\ ( enterprise) )
Ent-Host (______________________)
Figure 3: Secure Remote Access Server operation in an Enterprise
The following is an illustration of secure remote access data flow as
end-to-end IP packets traverse the Internet and the SRAS. The example
shows IP packet tunneling and IPsec transformation as packets are
exchanged between a remote Access host (RA-Host) and a host within
the enterprise (say, Ent-Host).
Note, the IP packets originating from or directed to RA-Host are
shown within PPP encapsulation, whereas, all other packets are shown
simply as IP packets. It is done this way to highlight the PPP
packets encapsulated within L2TP tunnel. The PPP headers below are
identified by their logical source and destination in parenthesis.
Note, however, the source and recipient information of the PPP data
is not a part of PPP header. This is described thus, just for
clarity. In the case of an L2TP tunnel, the L2TP header carries the
PPP session ID, which indirectly identifies the PPP end points to the
LAC and the LNS. Lastly, the IPsec Headers section below include the
tunneling overhead and the AH/ESP headers that are attached to the
tunnel.
RA-Host to Ent-Host Packet traversal:
------------------------------------
RA-Host LAC SRAS Ent-Host
=====================================================================
+----------------------+
PPP Header
(RA-Host ->SRAS)
+----------------------+
Tunnel-Mode IPsec
Hdr(s)(RA-Host->SRAS)
+----------------------+
End-to-end IP packet
transformed as needed
(RA-Host->Ent-Host)
+----------------------+
---------------------->
+----------------------+
IP Header
(LAC->SRAS)
+----------------------+
UDP Header
+----------------------+
L2TP Header
(incl. PPP Sess-ID)
+----------------------+
PPP Header
(RA-Host ->SRAS)
+----------------------+
Tunnel-Mode IPsec
Hdr(s)(RA-Host->SRAS)
+----------------------+
End-to-end IP packet
transformed as needed
(RA-Host->Ent-Host)
+----------------------+
---------------------->
+----------------------+
End-to-end IP packet
(RA-Host->Ent-Host)
+----------------------+
---------------------->
Ent-Host to RA-Host Packet traversal:
------------------------------------
Ent-Host SRAS LAC RA-Host
=====================================================================
+----------------------+
End-to-end IP packet
(Ent-Host->Ra-Host)
+----------------------+
---------------------->
+----------------------+
IP Header
(SRAS->LAC)
+----------------------+
UDP Header
+----------------------+
L2TP Header
(incl. PPP Sess-ID)
+----------------------+
PPP Header
(SRAS->RA-Host)
+----------------------+
Tunnel-Mode IPsec
Hdr(s)(SRAS->RA-Host)
+----------------------+
End-to-end IP packet
transformed as needed
(Ent-Host->RA-Host)
+----------------------+
---------------------->
+----------------------+
PPP Header
(SRAS->RA-Host)
+----------------------+
Tunnel-Mode IPsec
Hdr(s)(SRAS->RA-Host)
+----------------------+
End-to-end IP packet
transformed as needed
(Ent-Host->RA-Host)
+----------------------+
---------------------->
6. Limitations to Secure Remote Access using L2TP
The SRAS model described is not without its limitations. Below is a
list of the limitations.
1. Tunneling overhead: There is considerable tunneling overhead on
the end-to-end IP packet. Arguably, there is overlap of
information between tunneling headers. This overhead will undercut
packet throughput.
The overhead is particularly apparent at the LAC and SRAS nodes.
Specifically, the SRAS has the additional computational overhead
of IPsec processing on all IP packets exchanged with remote users.
This can be a significant bottleneck in the ability of SRAS to
scale for large numbers of remote users.
2. Fragmentation and reassembly: Large IP packets may be required to
undergo Fragmentation and reassembly at the LAC or the LNS as a
result of multiple tunnel overhead tagged to the packet.
Fragmentation and reassembly can havoc on packet throughput and
latency. However, it is possible to avoid the overhead by reducing
the MTU permitted within PPP frames.
3. Multiple identity and authentication requirement: Remote Access
users are required to authenticate themselves to the SRAS in order
to be oBTain access to the link. Further, when they require the
use of IKE to automate IPsec key exchange, they will need to
authenticate once again with the same or different ID and a
distinct authentication approach. The authentication requirements
of IKE phase 1 [Ref 8] and LCP [Ref 3] are different.
However, it is possible to have a single authentication approach
(i.e., a single ID and authentication mechanism) that can be
shared between LCP and IKE phase 1. The Extended Authentication
Protocol(EAP) [Ref 4] may be used as the base to transport IKE
authentication mechanism into PPP. Note, the configuration
overhead is not a drag on the functionality perse.
4. Weak security of Link level authentication: As LCP packets
traverse the Internet, the Identity of the remote user and the
password (if a password is used) is sent in the clear. This makes
it a target for someone on the net to steal the information and
masquerade as remote user. Note, however, this type of password
stealing will not jeopardize the security of the enterprise per
se, but could result in denial of service to remote users. An
intruder can collect the password data and simply steal the link,
but will not be able to run any IP applications subsequently, as
the SRAS will fail non-IPsec packet data.
A better approach would be to employ Extended Authentication
Protocol (EAP) [Ref 4] and select an authentication technique that
is not prone to stealing over the Internet. Alternately, the LAC
and the SRAS may be independently configured to use IPsec to
secure all LCP traffic exchanged between themselves.
7. Configuring RADIUS to support Secure Remote Access.
A centralized RADIUS database is used by enterprises to maintain the
authentication and authorization requirements of the dial-in Users.
It is also believed that direct dial-in access (e.g., through the
PSTN network is) safe and trusted and does not need any scrutiny
outside of the link level authentication enforced in LCP. This belief
is certainly not shared with the dial-in access through the Internet.
So, while the same RADIUS database may be used for a user directly
dialing-in or dialing in through the Internet, the security
requirements may vary. The following RADIUS attributes may be used to
mandate IPsec for the users dialing-in through the Internet. The
exact values for the attributes and its values may be obtained from
IANA (refer Section 10).
7.1. Security mandate based on access method
A new RADIUS attribute IPSEC_MANDATE (91) may be defined for each
user. This attribute may be given one of the following values.
NONE (=0) No IPsec mandated on the IP packets
embedded within PPP.
LNS_AS_SRAS (=1) Mandates Tunnel mode IPsec on the IP
packets embedded within PPP, only so
long as the PPP session terminates
at an LNS. LNS would be the tunnel
mode IPsec end point.
SRAS (=2) Mandates Tunnel mode IPsec on the IP
packets embedded within PPP,
irrespective of the NAS type the PPP
terminates in. I.e., the IPsec mandate
is not specific to LNS alone, and is
applicable to any NAS, terminating
PPP. NAS would be the tunnel mode
IPsec end point.
When IPSEC_MANDATE attribute is set to one of LNS_AS_SRAS or SRAS,
that would direct the NAS to drop any IP packets in PPP that are not
associated with an AH or ESP protocol. As an exception, the NAS will
continue to process IKE packets (UDP packets, with source and
destination port set to 500) directed from remote users. Further, the
security profile parameter, defined in the following section may add
additional criteria for which security is not mandatory.
7.2. Security profile for the user
A new SECURITY_PROFILE (92) parameter may be defined in RADIUS to
describe security access requirements for the users. The profile
could contain information such as the access control security
filters, security preferences and the nature of Keys (manual or
automatic generated via the IKE protocol) used for security purposes.
The SECURITY-PROFILE attribute can be assigned a filename, as a
string of characters. The contents of the file could be vendor
specific. But, the contents should include (a) a prioritized list
access control security policies, (b) Security Association security
preferences associated with each security policy.
7.3. IKE negotiation profile for the user
If the security profile of a user requires dynamic generation of
security keys, the parameters necessary for IKE negotiation may be
configured separately using a new IKE_NEGOTIATION_PROFILE (93)
parameter in RADIUS. IKE-NEGOTIATION_PROFILE attribute may be
assigned a filename, as a string of characters. The contents of the
file could however be vendor specific. The contents would typically
include (a) the IKE ID of the user and SRAS, (b) preferred
authentication approach and the associated parameters, such as a
pre-shared-key or a pointer to X.509 digital Certificate, and, (c)
ISAKMP security negotiation preferences for phase I.
8. Acknowledgements
The author would like to eXPress sincere thanks to Steve Willens for
initially suggesting this idea. The author is also thankful to Steve
for the many informal conversations which were instrumental in the
author being able to appreciate the diverse needs of the Remote
Access area.
9. Security Considerations
This document is about providing secure remote access to enterprises
via the Internet. However, the document does not address security
issues for network layers other than IP. While the document focus is
on security over the Internet, the security model provided is not
limited to the Internet or the IP infrastructure alone. It may also
be applied over other transport media such as Frame Relay and ATM
clouds. If the transport media is a trusted private network
infrastructure, the security measures described may not be as much of
an issue. The solution suggested in the document is keeping in view
the trust model between a remote user and enterprise.
10. IANA Considerations
This document proposes a total of three new RADIUS attributes to be
maintained by the IANA. These attributes IPSEC_MANDATE,
SECURITY_PROFILE and IKE_NEGOTIATION_PROFILE may be assigned the
values 91, 92 and 93 respectively so as not to conflict with the
definitions for recognized radius types, as defined in
http://www.isi.edu/in-notes/iana/assignments/radius-types.
The following sub-section explains the criteria to be used by the
IANA to assign additional numbers as values to the IPSEC-MANDATE
attribute described in section 7.1.
10.1. IPSEC-MANDATE attribute Value
Values 0-2 of the IPSEC-MANDATE-Type Attribute are defined in Section
7.1; the remaining values [3-255] are available for assignment by the
IANA with IETF Consensus [Ref 11].
REFERENCES
[1] Townsley, W., Valencia, A., Rubens, A., Pall, G., Zorn, G. and
B. Palter, "Layer Two Tunneling Protocol L2TP", RFC2661, August
1999.
[2] Rigney, C., Rubens, A., Simpson, W. and S. Willens, "Remote
Authentication Dial In User Service (RADIUS)", RFC2138, April
1997.
[3] Simpson, W., "The Point-to-Point Protocol (PPP)", STD 51, RFC
1661, July 1994.
[4] Blunk, L. and Vollbrecht, J. "PPP Extensible Authentication
Protocol (EAP)", RFC2284, March 1998.
[5] Kent, S. and R. Atkinson, "Security Architecture for the
Internet Protocol", RFC2401, November 1998.
[6] Kent, S. and R. Atkinson, "IP Encapsulating Security Payload
(ESP)", RFC2406, November 1998.
[7] Kent, S. and R. Atkinson, "IP Authentication Header", RFC2402,
November 1998.
[8] Harkins, D. and D. Carrel, "The Internet Key Exchange (IKE)",
RFC2409, November 1998.
[9] Piper, D., "The Internet IP Security Domain of Interpretation
for ISAKMP", RFC2407, November 1998.
[10] Reynolds, J. and J. Postel, "Assigned Numbers", STD 2, RFC1700,
October 1994.
See also http://www.iana.org/numbers.Html
[11] Narten, T. and H. Alvestrand, "Guidelines for writing an IANA
Considerations Section in RFCs", BCP 26, RFC2434, October 1998.
[12] Meyer, G., "The PPP Encryption Control Protocol (ECP)", RFC
1968, June 1996.
[13] Sklower, K. and G. Meyer, "The PPP DES Encryption Protocol,
Version 2 (DESE-bis)", RFC2419, September 1998.
Author's Address
Pyda Srisuresh
Campio Communications
630 Alder Drive
Milpitas, CA 95035
U.S.A.
Phone: +1 (408) 519-3849
EMail: srisuresh@yahoo.com
Full Copyright Statement
Copyright (C) The Internet Society (2000). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Acknowledgement
Funding for the RFCEditor function is currently provided by the
Internet Society.
