If you’ve been in IT for more than a few years, you’ve experienced The Next Big Thing in the industry. Several times. If you’ve been around since 2010 you, no doubt, heard endless chats about The Cloud. Other culprits of the past include Big Data, the Internet of Things, DevOps … and today we’re adding SD-WAN into the Buzzword Bingo game.
Going to manufacturer and IT support websites doesn’t do a whole lot of good either. You find yourself futilely trying to decipher sentences like these:
“Holistically managed network backend via internal or ‘organic’ sources” or “globally productivate and future-proof network processes”*
*These phrases may have been exaggerated for dramatic effect.
Adding to the confusion is the fact that there are dozens of companies bringing SD-WAN to market.
Just Cut the Crap, What’s SD-WAN?
So what is SD-WAN? It’s shorthand for “software-defined wide area network.” Simply put, SD-WAN is a technology that decouples the WAN from the physical infrastructure. In a matter of speaking, you are virtualizing the WAN infrastructure. Don’t call it that, though, because talking about “VWAN” on a crowded conference call is bound to grind the entire meeting to a halt.
What Does Using SD-WAN Actually Look Like?
When you go with an SD-WAN solution, you are provided either physical hardware or a virtual appliance for each site. The purpose of this appliance is to terminate all of your connections. Some vendors will only accept direct internet access (DIA) connections while others will accept anything you give it. This appliance will then be your next hop out of the network. Based on your configuration, the intelligence inside the box will determine which physical path will be used for your traffic.
You can configure traffic paths at the application layer. This will allow you to reserve the low latency, faster connection for VoIP traffic, and so on. You can also configure failover paths, QoS, firewall policies, UTM policies, depending on your product and configuration. Also, depending on the particular solution, you can place VPNs side by side with traditional WAN technologies like MPLS. I can hear some of you muttering “single point of failure.” Fear not, you can have your SD-WAN hardware arranged in a highly available (HA) configuration. Additionally, many come with ‘zero touch deployments’ that allow you to pre-deploy your SD-WAN anywhere around the globe. You just need someone to plug it in. That makes the cold spare strategy a lot easier to live with.
- Replace costly WAN circuits (depending on the cost of Internet circuits and the SD-WAN product)
- Remove complicated network configurations. You can dramatically reduce your dependency on routing protocols, IP SLAs, policy-based routes and so on.
- Reduce the need for high-end network engineering talent (you can just ask your favorite Mirazon engineer to do that for you as needed!)
- Add a high level of redundancy to your network with little effort.
- Add the ability for your network to adjust for poor conditions with little effort.
- Can be expensive if you are not careful, thus defeating the purpose.
- You are often dependent on Internet circuits which may rate lower than dedicated circuits on performance.
Here is what you need to think about when considering an SD-WAN solution:
- Does the billing model work for you? There are SD-WAN vendors that are up-front capital expenses and some that are recurring operational expenses.
- Is this going to be cost effective? Look for real numbers to compare against your current WAN infrastructure. Also, consider what renegotiating your circuits or switching to a new provider(s) will cost.
- Why are you considering SD-WAN? Lowering your bills? Increasing redundancy? Increasing performance? Simplifying your network architecture?