The OSI networking model, not unlike a good burrito, has seven layers. And just like that, the analogy fails. Where each layer of the burrito is experienced and savored, a network engineer might only see layers 2 and 3 of the OSI model, leaving the remaining five layers in anonymity.
This blog intends, in brevity, to demystify the two commonly used OSI layers and provide a practical guide on when to pick between layer 2 and layer 3 for remote and distributed applications.
First, an explanation.
Layer 2, known as the Data Link Layer, provides node-to-node data transfer with MAC address identification. All nodes on a layer 2 network are visible to one another. Ethernet switches are a common layer 2 example.
Layer 3, known as the Network Layer routes data packets to specific nodes identified by IP addresses. Visibility amongst nodes is point-to-point. As its function suggests, routers are a common layer 3 example.
Both have merits for different applications.
Layer 2 is beneficial for remote support. Enabling both the full functionality of most industrial protocols that sit atop ethernet and the ability to search for active IP nodes puts the programmer next to the machine, albeit virtually. There is no functional difference in this example between a layer 2 network and an ethernet cord.
Layer 3 provides a more scalable network infrastructure, like that used in a globally distributed system of machines connected to a central cloud for data analytics. Because data packets are routed, layer 3 enables 1-1 NAT and VLAN, among other traditional networking strategies. As distributed networks scale, layer 3 becomes more practical for managing connections.
If you would like more information on how to pick the right layer for your application, check our helpdesk article.
Blog post is written by Skylar Dhaese - OT Network engineer helping integrators and manufacturers monetize their digital transformation.