Category Archives: Juniper

OSPF Type-4 LSA

Have you ever questioned your self, why on earth do we need OSPF type-4 LSA ? well, you are not the only one. Welcome to this blog post where we’ll try to discover the actual worth of that LSA.

First let’s examine the below graph, where we have a basic implementation of multi-area OSPF. R1 RID is 1.1.1.1, R2 RID is 2.2.2.2, R3 RID is 3.3.3.3, and we have network 200.200.200.0.24 redistributed to OSPF area 1 by area 1 ASBR; R1.

Capture

We’ll start by examining the contents of the Link State Data Base (LSDB) of R3 for the external prefix 200.200.200/24;

Capture

Capture

When we look extensively at the type-5 LSA of prefix 200.200.200/24 in the LSDB of R3, we can see that it has an advertising router of 1.1.1.1 and a FW addr: 0.0.0.0.

The 1.1.1.1 is the RID of the advertising router, while the forwarding address is a way to that can be used to route traffic in another direction than to the router that originated the LSA. Having a value of 0.0.0.0 as the FW addr essentially tells other routers to forward packets for that destination to the router that advertised the LSA.

However when looking closely, we can see that R3 LSDB also has a type-3 LSA for 1.1.1.1/32,  so the question that popped out of my mind was “If there is a valid reachable type-3 for 1.1.1.1/32, then why do I need Type-4 LSA?”.

After taking a while looking into the issue I realized what went wrong. Well, I made a mistake, the 1.1.1.1/32 summary LSA was not the same 1.1.1.1 that advertised the external LSA. So what is the difference ?

Here’s were everything starts to get clear, we all network engineers have the convention that OSPF RID is driven from the highest loopback interface IP address present on the router and in case of the absence of loopback interfaces, the highest physical interface IP address well be chosen as the OSPF RID of the router, here is the catch, OSPF RID is a node-identifier. It usually uses an IPv4 address format, but it is an ID and does not need to match any interface present on the router.

Ok then, let’s try something, we’ll change the IP address of the lo0.0 of R1 to 100.100.100.100/32 and see what happens.

Capture

If we examined the external LSA of prefix 200.200.200.0/24 we notice that it still has the 1.1.1.1 as an advertising router, but wait, where is the LSA type-3 that was advertising 1.1.1.1/32, it’s no longer there and was replaced with 100.100.100.100/32 that is the new lo0.0 address of R1 that was already included under protocol ospf hierarchy of the configuration stanza. So the only way to reach the RID of R1 1.1.1.1 and consequently the external prefix 200.200.200/24 is by following the information contained in the ASBRSum LSA.

As you notice, our small lab was fruitful and we can conclude the following:

  • Routers in other areas see a Type-5 with a meaningless RID. They might have a type-3 prefix that looks like the ASBRs RID, but this doesn’t help it find the ASBR.
  • The type-4 LSA generated by the ABR is needed as a glue record, so that routers can ‘anchor’ the unknown ASBR RID against as reachable via a well-known node, the local ABR.

I hope this was informative and thank you for viewing.

Trouble Shooting OSPF Adjacency Problems (3)

In today’s post we are going to resume our discussion about the reason that could lead to OSPF adjacency problems, and we’ll look into the below reasons:

  • Mismatched hello and dead interval values.
  • Mismatched MTU settings.

We’ll work on the same topology as below.

capture2

Read the rest of this entry

Trouble Shooting OSPF Adjacency Problems (2)

In the last post we’ve discussed some of the reasons that could lead to OSPF adjacency problems, in today’s post we are going to look at the below reasons and we shall continue the rest of the reasons in a subsequent post.

  • Mismatched interface types.
  • OSPF priority is set to 0 on both sides.
  • Mismatched area IDs or mismatched area types.

We’ll continue to work on the same topology as below.

capture2

Read the rest of this entry

Trouble Shooting OSPF Adjacency Problems (1)

In this topic we are going to discuss some of the reasons that affect the OSPF adjacency between two peers and how to trouble shoot these issues in Junos.

First we will list the possible reasons for OSPF adjacency issues and we’ll discuss it in detail in subsequent paragraphs.

Possible Causes of OSPF Adjacency Issues:

  • Duplicate RIDs.
  • Mismatched subnet masks, or incorrect IP addressing.
  • Authentication mismatches.
  • Mismatched interface types.
  • OSPF priority is set to 0 on both sides.
  • Mismatched area IDs or mismatched area types.
  • Mismatched hello and dead interval values.
  • Mismatched MTU settings.

Read the rest of this entry

Juniper Aggregated Ethernet Interfaces

Overview

In this blog, we’ll be discussing one of the most High Availability protocols that’s broadly used in today’s networks that is Link Aggregation Control Protocol (LACP) and Link Aggregation Groups (ALG)

LACP is a method of bundling several physical interfaces to form one logical interface, which is advantageous in providing more bandwidth and increasing redundancy, On Cisco devices this is referred to as Ether-Channel.

It’s worth noting that the load-balancing hash algorithm for IP traffic uses criteria at Layer 2, Layer 3, and Layer 4. No configuration is necessary to enable load balancing. The load-balancing hash algorithm for non-IP traffic uses source and destination MAC addresses.

LACP exchanges are made between actors and partners. An actor is the local interface in an LACP exchange. A partner is the remote interface in an LACP exchange. LACP is defined in IEEE 802.3ad,  Aggregation of Multiple Link Segments and was designed to achieve the following:

  • Automatic addition and deletion of individual links to the aggregate bundle without user intervention
  • Link monitoring to check whether both ends of the bundle are connected to the correct group

Note that the Junos OS implementation of LACP provides link monitoring but not automatic addition and deletion of links.

The LACP mode can be active or passive. By default, when LACP is configured its mode defaults to the passive mode on aggregated Ethernet interfaces. To initiate transmission of LACP packets and response to LACP packets, you must enable LACP active mode.

Note that LACP exchanges protocol data units (PDUs) across all member links to ensure each physical interfaces is configured and functioning properly. Read the rest of this entry

L2 And L3 VPN over Ethernet Ring

In this post we are going to discuss how can we setup L2VPN and L3VPN over an Ethernet ring network, this is very challenging actually, knowing the fact that we need to achieve redundancy for both our edge network and to our customer.

As illustrated below, we have our MX104 PE router connected to the ring network that’s connected to MSAN cabins (Access Layer) through 2 Gigabit interfaces in order to achieve redundancy over the network.

1

The 1st challenge to take is how to setup L3VPN.

In our situaion we can’t use a logical unit of one of the main Gigs as this will not achieve the needed redundancy for the customer as there’s one Gig that will be up and the other will be down in order to prevent loops. So  the right solution would be to use Integrated Routing & Bridging interfaces (IRB) and assign it to the same bridge domain as the logical units of the main Gigs.

root@R02J> show interfaces descriptions | match Cust-A
ge-0/0/1.700  up     up    VPN: Cust-A –> Main Gig the customer is working on right now
ge-0/1/1.700   up    down VPN: Cust-A –> Backup Gig for the same customer
irb.700             up     up       VPN: Cust-A –>Layer 3 IRB interface for this customer (like an SVI interface on a Cisco Switches)

Read the rest of this entry