Layered Up: OSI Model Explained Through WAN and Connectivity

In this post, we delve into the OSI model and utilize connectivity options in each of the first three OSI layers to help explain how the model works

data center osi model
Matt Pinto

Aug 31, 2022


Determining exactly what you’re getting when talking to a service provider can be an exercise in frustration. They’re likely to use technical jargon, and you might find yourself nodding with a forced smile as they ramble on with jargon about OSI layers, security, and how your data gets routed. 

Likelier still is that you’re staring at this information in a customer service email and wishing someone could just explain your options in clear, concise terms.

We totally get it. So, in this post, we delve into the OSI model and utilize connectivity options in each of the first three layers to help explain how the model works. 

The OSI Model

A quick refresher: theOSI (Open Systems Interconnection) Model is a way of classifying networks across seven different layers, with each incremental layer abstracting the user further and further away from so-called “bits on a wire.” The OSI layers we’re particularly interested in are the first three:

  1. The Physical layer refers to hardware: fiber-optic cables, repeaters, and modems (basically the physical-world infrastructure that makes up your network).

  2. The Data Link layer refers to communications between directly connected nodes. If you’re old enough to remember the 90s, think of this as the information being sent between PCs during a LAN party (and if you’re not that old,here you go).

  3. The Network Layer refers to routed connections, where data is transferred between networks. For example, messages sent through your email are routed to their destination via this layer. 

In the Beginning, There Was Darkness

With the definitions out of the way, let’s look at what’s on offer in each case:

Regarding connectivity services and Wide Area Networks (WANs), there really isn’t much available to rent in the Physical Layer. The main option is dark fiber, i.e., physical fiber infrastructure that a network operator is required to “light up” for data transmission. In this case, “dark” = no data transmission = “not lit.” As soon as any data transmission occurs, we’re technically in Layer 2 of the OSI model, as you’ve now taken a physical link and created connected nodes by lighting the fiber to transmit data from an A to a Z location. 

While many companies probably wouldn’t have a use for dark fiber (it’s easier to buy lit fiber services), there are some advantages. For starters, you’re renting a high-capacity, ultra-secure connection with infinite room to scale your operations. (This could be very cool if you’re planning on starting a spy agency, for example.). Actually though, for the most data-intensive operations out there where a dark fiber lease is reasonably accessible, it is often the most economic option, so long as you have the networking intel on hand to operate the network. 

The downsides are that it’s tough and costly to procure dark fiber. Most carriers don’t want to rent it to you (because it means the physical asset is removed from their inventory). If they are willing to rent it out, you’ll probably get locked into a very long contract (think 3-20 years!). You’ll also have to install the relevant equipment to light it up and need a team capable of handling the technical aspects of keeping your network running. 

This option certainly isn’t our go-to recommendation for most businesses. But if the possibilities of dark fiber still intrigue you, check out our more detailed rundown of the pros and conshere. We also wrote a post on when dark fiber makes sense.

C’mon Baby, Light My Fiber!

When you reach the Data Link layer, network operations become easier, and there’s a little more on offer. Note that connections in this layer are still not routed via the internet, instead relying on direct connections between sites to transmit data.

Some popular network service options for business at this level are Point-to-Point (PTP), Wavelength Services (Wave), and Virtual Private LAN Services (VPLS). Unlike hard-to-procure dark fiber, these commodity services are more readily available through several providers.

  • PTP Connections (P2P or private line): This type of connection links two sites in a closed network. With a PTP connection, data always travels in the same way along a dedicated path between the two sites. The advantages of having this dedicated route are connections that are very secure and that typically offer decent bandwidth and low latency.

  • Wave: A similar option to PTP is Wave. While this alternative offers many of the same benefits, a significant difference here lies in how data is transmitted. With Wave, more data can be sent over the same strand of fiberbecause Wave uses a technique calledDense Wavelength Division Multiplexing (DWDM). While the name is a mouthful, the technique is straightforward: multiple data batches are sent simultaneously using different light wavelengths (the wavelengths don’t mix when transmitted together). So if your company has ultra-high bandwidth needs, Wave might be a much better fit than PTP in Layer 2. You can check out our guidehere to get more insight into how these options stack up.

  • VPLS: The last option to consider is a point-to-multipoint Virtual Private LAN Service that can hook up multiple sites in the same network. VPLS shares many similarities with Multiprotocol Label Switching (MPLS – discussed below), but the difference is that VPLS is Layer 2, while MPLS is a Layer 3 (routed) solution. And that means better security. 

Unfortunately, it also means that a VPLS network requires “full mesh” – in other words, every site needs to be directly connected to each other site, adding to the expense as a single carrier needs to provide those connections. Overall, this means the cost of ownership is prohibitive. And if the VPLS carrier network develops issues, your data isn’t going anywhere until they’ve fixed it. As with many premium offerings, you should have a robust Service Level Agreement (SLA) to encourage them to get things back up and running.

It's worth noting that many businesses also consider a combination of VPLS and MPLS to get the best of both worlds – an easily scalable network with secure channels for sensitive data.

A World of (Networked) Possibilities

If the idea of deploying your own internal network means you can’t readily go check out the latest cat memes, fear not! We’ve reached Layer 3 now, where networks route data between one another, and that means regular old internet access is an option (though not necessarily the best one).

Going with bog-standard, best-effort (broadband) internet is fine for many smaller companies. Many issues at this level center around customizability, which is very low in the case of broadband. It's also worth noting that you'll want to ensure the data is highly secure if you deal with sensitive customer data. Keeping things secure is more challenging when everything traverses the internet.

Data services on this end of the spectrum also tend to be asymmetrical, meaning your upload and download speeds will differ. This situation can be frustrating when you need to move large chunks of data quickly.

A more balanced option is a symmetrical line where you get the same upload and download speeds, usually over fiber or fixed wireless. In most cases, this involves procuringDedicated Internet Access (DIA), where your business doesn’t have to share bandwidth with other users. Of course, this also means your provider is on the hook for guaranteed speeds, and that there’s (hopefully) better technical support when things go wrong.

Regardless of the type of internet access you go for, you also have the option of deploying anSD-WAN (Software-Defined Wide Area Network), which provides more security while being cheaper than our last option –Multiprotocol Label Switching (MPLS).

  • MPLS: Just like internet access, MPLS provides routed connections between sites. The difference is that MPLS connections are partitioned off from the public internet. And, unlike internet routing, data packets follow fixed, optimized pathways based on the type of data being sent. Essentially, the data packets all have a “label” that predetermines where they go. So, in an MPLS circuit, voice and video data might be labeled to travel along the fastest possible route (reducing latency), whereas emails might take a slower path.

MPLS services typically use a single carrier's network (i.e., dedicated infrastructure). It also offers symmetrical bandwidth and provides stronger security than internet. Once again, the risk is that if the carrier has issues with its infrastructure, multiple sites will experience connectivity issues. However, as with most premium arrangements, they must still meet the obligations laid out in their SLA. The premium price tag does mean that it can be difficult and costly to scale up an MPLS network, making thechoice of MPLS or SD-WAN something to think over carefully.

Where to From Here?

The connectivity option that works best for you is determined by your business. If you’re still wondering: “where to go from here?”, we can help you sort through the possibilities. Just book a demo with us, and we’ll set you up with the perfect network solution.

Want to learn more about how Lightyear can help you?

Let us show you the product and discuss specifics on how it might be helpful.

Not ready to buy?

Stay up to date on our product, straight to your inbox every month.

Featured Articles