Cisco PPP Authentication

As a network engineer, you most probably already had to do with PPP authentication at least once or two times in your daily operation.  Even more, if you are going for a Cisco certification (and not only) you should know some stuff about PPP authentication. For today, I’ve planned to deal with back-to-back PPP authentication.

For this back to back scenario, we have the following simple topology:

When we talk about PPP authentication on a end-to-end line we are dealing with 3 major authentication method:

PAP

CHAP

EAP

Now, when we think at security, we can easily observe that PAP is the less secure one and CHAP or EAP are the strongest one.

PAP (Password Authetication Protocol) transmits unencrypted ASCII passwords over the network and is therefore considered insecure. It should be used only as a last resort when the remote server does not support a stronger authentication protocol, like CHAP or EAP.
CHAP (Challenge-Handshake Authentication Protocol) is a more secure protocol as it uses a three-way handshake and the shared secret (password) is never sent on the wires. Instead a MD5 hash checksum is calculated based on the share secret and this one is sent as a challenge to the other peer.
EAP (Extensible Authentication Protocol) is an authentication framework, not a specific authentication mechanism. It provides some common functions and negotiation of authentication methods called EAP methods. There are currently about 40 different methods defined.

When we think of PPP authentication direction there are 2 types:

– one way authentication

– two ways authentication

Pretty obvious, no?

OK, enough with the theory. If you need some more deep understanding of PPP, there is always Internet. Next, I will show you some simple example with PPP authetication using PAP, CHAP and EAP.

PAP type authentication

Let’s assume that in the scenario above, R1 is sending a challenge to R2. Very important! From PPP authentication configuration, you don’t have to do anything to response to a challenge. This is done automatically.

R1
username R2 password cisco
int s0/0
encapsulation ppp
ppp authentication pap

R2
int s0/0
encapsulation ppp
ppp pap sent-user R2 password cisco

Actually that’s it. As I said, pretty simple. You configure R1 to send an authentication challenge to R2 with “ppp authentication pap”. R2 has to reply to this challenge with a username and a password defined with “ppp pap sent-user R2 password cisco”. This username and password have to be defined on R1. No matter what user and password you define on challenged part to be sent back, that information you have to define on challenger.

CHAP authentication type

Configuring CHAP is even easier. In the example below, I will configure R2 to send a CHAP challenge to R1

R1
username R2 password cisco
int s0/0
encapsulation ppp

R2
username R1 password cisco
int s0/0
encapsulation ppp
ppp authentication chap

By default, CHAP is sending the router hostname the user in the three-way handshake process, so there is no need to specify what user to send like in PAP method. As I said before, this method is more secure than PAP.

EAP authetication type

To be honest I didn’t saw too many PPP connections being authenticated with EAP, but is there and I saw some CCIE lab topics so, you should keep an eye on it. Like the other two method this one is easy to implement and is offering more secure level of authetication than PAP.

R1
username R2 password cisco
int s0/0
encapsulation ppp
ppp authentication eap
ppp eap identity R1
ppp eap password cisco
ppp eap local

R2
username R1 password cisco
int s0/0
encapsulation ppp
ppp authentication eap
ppp eap identity R2
ppp eap password cisco
ppp eap local

I believe that the command syntax is telling pretty much all there is. With “identity” you define the user to be send to the peer, “password” it what word is saying and last option “local” is the quite important. By default EAP needs a RADIUS server for authentication. If you don’t have one (exams, quick testing…) then you want EAP to use local database instead of RADIUS.

This  are the basics of PPP authentication. Even if there are not so much in use, try to remember this small steps as you might need them sometimes.

SNMP Version 3 Authentication Vulnerabilities

Multiple Cisco products contain either of two authentication vulnerabilities in the Simple Network Management Protocol version 3 (SNMPv3) feature. These vulnerabilities can be exploited when processing a malformed SNMPv3 message. These vulnerabilities could allow the disclosure of network information or may enable an attacker to perform configuration changes to vulnerable devices. The SNMP server is an optional service that is disabled by default in Cisco products. Only SNMPv3 is impacted by these vulnerabilities. Workarounds are available for mitigating the impact of the vulnerabilities described in this document.

Vulnerable products:
# Cisco IOS
# Cisco IOS-XR
# Cisco Catalyst Operating System (CatOS)
# Cisco NX-OS
# Cisco Application Control Engine (ACE) Module
# Cisco ACE Appliance
# Cisco ACE XML Gateway
# Cisco MDS 9000 Series Multilayer Fabric Switches
# Cisco Wireless LAN Controller (WLC)
# Cisco Application and Content Networking System (ACNS)
# Cisco Wide Area Application Services (WAAS)
# Cisco MGX 8850, 8880 Media Gateway and Switch
# Cisco PSTN Gateway (PGW2200)

Read more on Cisco Security Advisory

Authentication with a twist

fujitsu_logoPeople mistrust fingerprint devices, mostly because they associate fingerprinting with criminal activity. The average citizen thinks that the fingerprint registration could be stolen and used to implicate them in a crime. It can’t, of course, but that doesn’t change their perception. The same problem faces facial scanning/recognition software which has been used (unsuccessfully) to identify wanted criminals at sporting events. Retina scanners simply scare people – they don’t want anything being shined into their eye. So what can we do?

The smart folks at Fujitsu have come up with a new system to read a biometric. It’s non-intrusive, isn’t likely to be featured at a crime scene on a TV series but does provide a unique signature with little effort on the user’s part.

Read the full article on NetworkWorld.com

Cisco: Multilink PPP over Frame Relay (MLPoFR)

In this tutorial I propose to show something that is not very used these days, or at least not every day, but which can be tricky if you don’t know how to approach this type of configuration. To understand this, I assume that you know the basics about PPP, FR and Multilink. I will make a short summary here but I will not go into details:

PPP or Point-to-Point protocol is used to establish direct connection between two network points. It can provide authentication, encryption privacy and compression.
FR or Frame-Relay is a telecommunication service used mostly on the WAN side towards your provider or carrier and it relay on frames for data transmission.
Multilink is used for bundle together 2 or more channels / circuits for communication improvement.

Here we will use these 3 technology to create something called MLPoFR. For security we will use authentication. Please download here the topology. Please be aware that in the topology you cannot see actually 2 links there (it a limitation of GNS3), but trust me the links are there. To be more convenient and quick the R1 of the topology is preconfigured.

Please see the tutorial below:

Cisco: Deny false information routing injection into OSPF domain

In a well controlled environment, false information routing should not reach your OSPF domain, as network engineer take care what to advertise and what not into OSPF. But there are cases when you have to deal with 3rd party companies somehow, and you want to be sure that nothing in injected by mistake into your domain. Also this can be a task for CCIE RS lab exam.

And since I specified that this can be an exam task, let take some “DO NOT USE” rule and we have to accomplish the task above without using the command “ip ospf authentication message-digest”. Download the used topology here. R1 from the topology is pre-configured. The OSPF timers have been reconfigured to hello 1 second and dead interval 5 seconds, not to wait “forever” until it rebuilds the adjacency.

Please see the tutorial below: