7 things that you should know about Layer 3

As Layer 3 routing becomes more widely used, it offers the ability to bring intelligence to the edge of the network – at each and every node. But what does that actually mean for the performance and management of network infrastructure? In this guide, we look at the key benefits of a Layer 3 distributed intelligence architecture.

Monday 23rd June 14

1) Climbing the stack isn't dumb

Layer 3 routing is the step on from Layer 2 switching. A Layer 2 switch works at the data layer (Layer 2) of the IP stack whereas a Layer 3 router works at the network layer (Layer3). When using Layer 2 switching, all broadcast data packets are forwarded throughout the network unless filters are applied. Most Layer 2 switches use MAC addressing to connect between devices but these have to be learnt over the whole network at the start of a session.  In the Layer 3 world, there is intercommunication between networks, routers and users both locally and at remote locations. This communications allows intelligence to be added to each router to ensure network traffic is more effectively routed.

2) Let your router decide

Apart form moving from switching to routing, Layer 3 allows network services such as traffic management and firewalls to be implemented at an individual router level. There is no need for a central controlling device as there would be in a Layer 2 network. In a Layer 2 network the firewall tends to be an edge/egress device. In Layer 3, it can be a set of distributed devices.

3) More data, less flooding

A major drawback for a Layer 2 network is that, in effect, every device on the network is polled when data packets are sent. This is a very inefficient use of bandwidth. One solution to this was to create Layer 2 VLANs to create a series of subnets to relieve some network congestion. The drawback is that the network still needed to route at a Layer 3 level between Layer 2 VLAN if they were to communicate. A Layer 3 network uses routing to ensure that the data packet goes directly to subnet or individual end port without polling any other device. This significantly reduces bandwidth congestion.

4) More red lights for unwanted traffic

Layer 3 routers allow granular traffic management to be set at an individual user, device and application level. Not only can finely tuned traffic priorities can be set, precision rate limiting, congestion management and interference avoidance protects against network incidents capable of bringing networks to a halt. This allows for greater 'goodput' – higher consistent data rates – across the network with reduced points of failure.

5) A step forward in security

By segregating traffic over the network, a Layer 3 architecture is inherently more secure.  By extending Layer 3 intelligence to the edge reduces the surface area that is available for attack. The reason for this is that Layer 2 switches have limited security capabilities with security policies set and executed by the central controller. Layer 3 allows granular security rules to be administered on the router via Layer 3 Access Control Lists. So if any router is compromised, the attacker only has access to the subnet controlled by that device.

6) Making your life easier

Configuration and management of a Layer 3 network is more straightforward. The network administrator can manage edge routers through web-based or SMTP access. What's more, many Layer 3 protocols are dynamic once implemented. A well designed and implemented Layer 3 network also allows the administrator to set up automatic self-healing routes for data to enable near real time recovery of communications should one device fail.
7) It's all about quality

An intelligent Layer 3 architecture allows the administrator to set stringent Quality of Service (QoS) policies. With policies administered on the individual edge routers, the network can be partitioned into multiple priority levels. The policies will determine priority for different classes of service or the network rights and application access of individual user or user group.

There are still many instances where Layer 2 is perfectly acceptable for network requirements. However, by implementing Layer 3 routing at the edge, the network will benefit from improved security, availability and bandwidth utilisation.