Protecting The Router Control Plane
draft-dugal-opsec-protect-control-plane-01.txt

Abstract

This memo provides a method for protecting a router's control plane from undesired or malicious traffic. In this approach, all legitimate control plane traffic is identifed. Once legitimate traffic has been identified, a filter is deployed on the router's forwarding plane. That filter prevents traffic not specifically identified as legitimate from reaching the router's control plane.

Status of this Memo

This Internet-Draft is submitted to IETF in full conformance with the provisions of BCP 78 and BCP 79.

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.

This Internet-Draft will expire on August 2, 2010.

Copyright Notice

Copyright (c) 2010 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 (http://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 BSD License.

Table of Contents

1.  Introduction
2.  Applicability Statement
3.  Method
    3.1.  Legitimate Traffic
    3.2.  Filter Design
    3.3.  Design Trade-offs
4.  Security Considerations
5.  IANA Considerations
6.  Acknowledgements
7.  Informative References
Appendix A.  Cisco Configuration
Appendix B.  Juniper Configuration
§  Authors' Addresses

1.  Introduction

Modern core router architecture design maintains a strict separation of forwarding and control plane hardware and software. Forwarding plane is typically described as router architecture hardware and software components responsible for taking a packet coming in one interface, performing a lookup to identify the packet's IP next hop and determine the outgoing interface towards the destination, and forwarding the packet through the correct outgoing interface. Control plane supports routing and management functions, and is generally described as as router architecture hardware and software components for handling packets destined to the device itself as well as building and sending packets originated locally on the device.

Visually it can be represented as the control plane hardware sitting on top of and interfacing with the forwarding plane hardware, with interfaces connecting to other network devices. See Figure 1 (Router Control Plane Protection).

              +---------------+
              | Control Plane |
              +------+ +------+
                     | |
                Router Control
               Plane Protection
                     | |
              +------+ +------+
              |   Forwarding  |
Interface X ==[     Plane     ]== Interface Y
              +---------------+

Typically, forwarding plane functionality is realized in high-performance Application Specific Integrated Circuits (ASICs) that are capable of handling very high packet rates. By contrast, the control plane is generally realized in software on general purpose processors. While software instructions run on both planes, the control plane software is usually not optimized for high speed packet handling. Given their differences in packet handling capabilities, control plane hardware is more suceptible to be overwhelmed by a DoS attack than forwarding plane ASICs. It is imperative that the control plane remain stable regardless of traffic load to and from the device because the control plane is what drives the programming of the forwarding plane.

The control plane processes traffic destined to the router, and because of the wider range of functionality is more suceptible to security vulnerabilities and a more likely target for a DoS attack than the forwarding plane.

It is advisable to protect the control plane by implementing mechanisms to filter completely or rate limit traffic not required at the control plane level (i.e., unwanted traffic). Control Plane Protection is the concept of filtering traffic unwanted traffic which would be diverted out of the forwarding plane up to the control plane. The closer to the forwarding plane and line-rate hardware the filters and rate-limiters are, the more effective the protection is and the more resistent the system is to DoS attacks. This memo demonstrates how to deploy such a filter.

2.  Applicability Statement

The method described in Section 3 (Method) is a sample illustration to demonstrate how to protect the router control plane from unwanted traffic. Recognizing that deployment scenarios will vary, the exact implementation is not generally applicable in all situations. The cateorization of legitimate control plane traffic is critically important in a successul implementation.

The examples given in this memo are simplified and minimalistic, designed to illustrate the concept of protecting the router's control plane. From them, operators can extrapolate specifics based on their unique configuration and environment.

This method is applicable for IPv4 as well as IPv6 traffic. The sample legitimate traffic in Section 3.1 (Legitimate Traffic) uses IPv4 addresses, but can be expanded to IPv6 as well.

3.  Method

In this memo, the authors demonstrate how a filter protecting the control plane can be deployed. In Section 3.1 (Legitimate Traffic), a sample router introduced and all traffic that its control plane must process is identified. In Section 3.2 (Filter Design), filter design concepts are discussed. Cisco (Cisco IOS software) and Juniper (JUNOS) implementations are provided in Appendices A (Cisco Configuration) and B (Juniper Configuration), respectively.

3.1.  Legitimate Traffic

In this example, the router control plane must process traffic from the following sources:

• ICMP traffic from any source, rate-limited to 2 Mbps
• OSPF traffic from routers within own network (network mask 192.0.2.0/24)
• iBGP traffic from routers within own network (network mask 192.0.2.0/24)
• eBGP traffic from known eBGP peers (network addresses 198.51.100.25, 198.51.100.27, 198.51.100.29, 198.51.100.31)
• DNS traffic from local DNS resolvers (198.51.100.0/30)
• NTP traffic from local NTP (198.51.100.4/30)
• SSH traffic from network management stations (198.51.100.128/25)
• SNMP traffic from network management stations (198.51.100.128/25)

The characteristics of legitimate traffic will vary from network to network. The list provided above is for example only.

3.2.  Filter Design

A filter is installed on the forwarding plane. This filter counts and silently discards all traffic not matching the profile provided in Section 3.1 (Legitimate Traffic). Because the filter is enforced on the forwarding plane, it prevents unwanted traffic from consuming bandwidth on the interface that connects the forwarding plane to the control plane. The counters serve as an important forensic tool for the analysis of potential attacks, and as an invaluable debugging and troubleshooting aid.

A rate limiter also is installed on the forwarding plane. The rate limiter restricts ICMP traffic bound for the control plane to some reasonable volume. In our example, we will rate limit to 2 Megabits per second (Mbps).

Syntactically, these filters explicitly define "allowed" traffic (including IP addresses, protocols, and ports), define acceptable actions for these acceptable traffic profiles (e.g., rate-limit or simply permit the traffic), and then drop to the bit bucket all traffic destined to the control plane but not explicitly allowed.

3.3.  Design Trade-offs

In designing the protection method, there are two independent parts to consider: the classification of traffic (i.e., which traffic is matched by the filters), and the policy actions taken on the classified traffic.

There are different levels of granularity utilized for traffic classification. For example, allowing all traffic from specific source IP addresses versus allowing only a specific set of protocols from those specific source IP addresses will each affect a different set of traffic.

Similarly, the policy actions taken on the classified traffic have degrees of impact that may not become immediately obvious. For example, discarding all ICMP traffic may have a negative impact on the operational use of ICMP tools such as ping or traceroute to debug network issues or to test turn up of a new circuit.

It is important to note that both classification and policy action decisions are accompanied by respective trade-offs. Two examples of these trade-off decisions are, operational complexity at the expense of policy (and statistics gathering) detail, and tighter protection at the expense of network supportability and troubleshooting ability.

The goal of the method for protecting the router control plane is to minimize potential disruptions. The granularity of the filter design inversely correlates to the scope of the potential disruption. The finer the granularity of the filter design (e.g., isolating kinds of sub-traffic from the rest of the policed traffic, or isolating valid source addresses into a different class or classes) the smaller the scope of disruption.

4.  Security Considerations

The filter above leaves the router susceptible to discovery from any host on the Internet. If network operators find this risk objectionable, they can mitigate it by restricting the sub-networks from which ICMP Echo requests are accepted.

The filter above also leaves the router exposed to port scans from hosts spoofing the source addresses found in Section 3.1 (Legitimate Traffic). Network operators can mitigate this risk by preventing source address spoofing with filters applied at the network edge. Refer to Section 5.3.8 of [RFC1812] (Baker, F., “Requirements for IP Version 4 Routers,” June 1995.) for more information regarding source address validation. Other methods also exist for limiting exposure to packet spoofing such as the Generalized TTL Security Mechanism (GTSM) [RFC5082] (Gill, V., Heasley, J., Meyer, D., Savola, P., and C. Pignataro, “The Generalized TTL Security Mechanism (GTSM),” October 2007.) and Ingress Filtering [RFC3704] (Baker, F. and P. Savola, “Ingress Filtering for Multihomed Networks,” March 2004.).

The ICMP rate limiter specified in this filter protects the router from floods of ICMP traffic. However, during an ICMP flood, some legitimate ICMP traffic may be dropped. Because of this, when operators discover a flood of ICMP traffic, they are highly motivated to cut it off at its source.

Additional considerations pertaining to the usage and handling of traffic that utilizes the IP Router Alter Options can be found at [I‑D.rahman‑rtg‑router‑alert‑considerations] (Faucheur, F., “IP Router Alert Considerations and Usage,” October 2009.).

The treatment of exception traffic in the forwarding plane, and the generation of specific messages by the control plane also requires protection from a DoS attack. Specifically, the generation of ICMP Unreachable messages by the control plane needs to be rate-limited, either implicitly within the router's architecture or explicitly through configuration. See Section 4.3.2.8 of [RFC1812] (Baker, F., “Requirements for IP Version 4 Routers,” June 1995.).

Additionally, the handling of TTL / Hop Limit expired traffic needs protection. For example, rate limiting the TTL / Hop Limit expired traffic before sending the packets to the control plane component that will send the ICMP error, and distributing the sending of ICMP errors in a Line Card CPU are protection mechanisms that deter attacks before a rate limited in the main control plane component.

5.  IANA Considerations

[RFC Editor: please remove this section prior to publication.]

This document has no IANA actions.

6.  Acknowledgements

The authors would like to thank Ron Bonica for providing review, suggestions, and valuable input. Pekka Savola, Warren Kumari and Xu Chen provided very thorough and useful feedback that improved the document.

7. Informative References

Appendix A.  Cisco Configuration

!Start: Protecting The Router Control Plane
!
!Policy-map Configuration
!
!Access-list Definitions
ip access-list extended ICMP
 permit icmp any any
ip access-list extended OSPF
 permit ospf 192.0.2.0 0.0.0.255 any
ip access-list extended IBGP
 permit tcp 192.0.2.0 0.0.0.255 eq bgp any
 permit tcp 192.0.2.0 0.0.0.255 any eq bgp
ip access-list extended EBGP
 permit tcp host 198.51.100.25 eq bgp any
 permit tcp host 198.51.100.25 any eq bgp
 permit tcp host 198.51.100.27 eq bgp any
 permit tcp host 198.51.100.27 any eq bgp
 permit tcp host 198.51.100.29 eq bgp any
 permit tcp host 198.51.100.29 any eq bgp
 permit tcp host 198.51.100.31 eq bgp any
 permit tcp host 198.51.100.31 any eq bgp
ip access-list extended DNS
 permit udp 198.51.100.0 0.0.0.252 eq domain any
ip access-list extended NTP
 permit udp 198.51.100.4 255.255.255.252 any eq ntp
ip access-list extended SSH
 permit tcp 198.51.100.0 0.0.0.128 any eq 22
ip access-list extended SNMP
 permit udp 198.51.100.128 0.0.0.125 eq snmp any
 permit udp 198.51.100.128 0.0.0.125 eq snmptrap any
!
!Class Definitions
!
class-map match-all ICMP
 match access-group name ICMP
class-map match-all OSPF
 match access-group name OSPF
class-map match-all IBGP
 match access-group name IBGP
class-map match-all EBGP
 match access-group name EBGP
class-map match-all DNS
 match access-group name DNS
class-map match-all NTP
 match access-group name NTP
class-map match-all SSH
 match access-group name SSH
class-map match-all SNMP
 match access-group name SNMP
!
!Policy Definition
!
policy-map COPP
 class ICMP
  police 2000000
 class OSPF
 class IBGP
 class EBGP
 class DNS
 class NTP
 class SSH
 class SNMP
 class class-default
   police cir 8000
     conform-action drop
     exceed-action drop
     violate-action drop
!
!Control Plane Configuration
!
control-plane
 service-policy input COPP
!
!End: Protecting The Router Control Plane

Appendix B.  Juniper Configuration

policy-options {
    prefix-list IBGP-NEIGHBORS {
        192.0.2.0/24;
    }
    prefix-list EBGP-NEIGHBORS {
        198.51.100.25/32;
        198.51.100.27/32;
        198.51.100.29/32;
        198.51.100.31/32;
    }
}
firewall {
    policer 2Mbps {
        if-exceeding {
            bandwidth-limit 2m;
            burst-size-limit 2k;
        }
        then discard;
    }
    family inet {
        filter protect-control-plane {
            term icmp {
                from {
                    protocol icmp;
                }
                policer 2Mbps;
                then accept;
            }
            term ospf {
                from {
                    source-address {
                        192.0.2.0/24;
                    }
                    protocol ospf;
                }
                then accept;
            }
            term ibgp-connect {
                from {
                    source-prefix-list {
                        IBGP-NEIGHBORS;
                    }
                    protocol tcp;
                    destination-port bgp;
                }
                then accept;
            }
           term ibgp-reply {
                from {
                    source-prefix-list {
                        IBGP-NEIGHBORS;
                    }
                    protocol tcp;
                    port bgp;
                }
                then accept;
            }
            term ebgp-connect {
                from {
                    source-prefix-list {
                        EBGP-NEIGHBORS;
                    }
                    protocol tcp;
                    destination-port bgp;
                }
                then accept;
            }
            term ebgp-reply {
                from {
                    source-prefix-list {
                        EBGP-NEIGHBORS;
                    }
                    protocol tcp;
                    port bgp;
                }
                then accept;
            }
            term dns {
                from {
                    source-address {
                        198.51.100.0/30;
                    }
                    protocol udp;
                    port domain;
                }
                then accept;
            }
            term ntp {
                from {
                    source-address {
                        198.51.100.4/30;
                    }
                    protocol udp;
                    destination-port ntp;
                }
                then accept;
            }
            term ssh {
                from {
                    source-address {
                        198.51.100.128/25;
                    }
                    protocol tcp;
                    destination-port ssh;
                }
                then accept;
            }
            term snmp {
                from {
                    source-address {
                        198.51.100.128/25;
                    }
                    protocol udp;
                    port [snmp snmptrap];
                }
                then accept;
            }
            term default-term {
                then {
                    count copp-discards;
                    log;
                    discard;
                }
            }
        }
    }
}
interfaces {
    lo0 {
        unit 0 {
            family inet {
                filter input protect-control-plane;
            }
         }
    }
}

Authors' Addresses