128 Technology (128T) has built a clever software router that takes a unique approach to getting your traffic from source to destination. Will anyone care?
That’s the question I’ve been asking myself since I attended a Networking Field Day event with 128T in July 2019.
At the event, I learned 128T calls its product the Session Smart Router. Here are the basic characteristics:
- The router runs on x86 hardware, and will scale as you give it more cores; no need for expensive ASICs
- The router makes efficient use of limited bandwidth
- It incorporates an L4 firewall and operates in default-deny out of the box
- It encrypts every packet’s payload
- You route based on sessions (that is, an application transaction) instead of packet-by-packet
- You can use multiple links at a site and send sessions across different links based on application type, link performance, and other business criteria
- You can apply performance and security policies on a session-by-session basis
- The router doesn’t rely on tunnels or encapsulation, which benefits overall performance
- A software controller, called the Conductor, tracks the configuration of each router and provides a global management platform, but does not sit in the data plane
For more technical details on how it works, see Ethan Banks’ excellent blog or 128 Technology’s Networking Field Day videos.
Difficult Positioning
If you were paying attention to the bullet points above, you might say to yourself “Hey, 128T’s router sounds a lot like SD-WAN.”
You are correct. 128T’s Session Smart Router is a very capable SD-WAN product and has all the key features and capabilities you’d want.
In addition to the features listed above, the Session Smart Router includes application identification, the ability to send different sessions across different link types based on policy, and the ability to switch sessions from one link to another based on link performance.
However, 128T doesn’t just want to sell you SD-WAN. They want to sell you a router that happens to do SD-WAN, as well as encryption and L4 firewalling.
Those are good features and capabilities, but they aren’t unique to 128T. So again, the question returns to how 128T is different from other routers, and whether those differences are significant enough to matter.
To answer that question, you really have to dive into 128T’s solution, often to excruciating detail.
If you’ll bear with me, here’s a little story to frame 128T’s problem.
Chord Changes
I’ve been playing the guitar for about thirty years. In that time I’ve developed deep-set muscle memory for how to form chords.
A couple of years ago my wife got me guitar lessons as a birthday present. I was excited to fill in a few of the vast gaps in my abilities.
At the first lesson, my instructor evaluated my playing. When he saw me make a D chord, he winced.
“Try it like this,” he said. It was a small change in finger positioning, and he said it would make for more efficient chord changes and chord variations.
I tried. I could make the shape, but not naturally. My fingers were clumsy, and my hand resisted. It wanted to go back to shape it had been making for 30 years. I had to stop mid-song to think about the strings and frets and where the index finger should go in relation to the ring and middle fingers. It sucked.
After a few frustrating minutes, I had a conversation with myself. Would the benefits of this new D chord vastly outweigh the benefits of my existing D chord? No.
Would the practice necessary to learn this new formation cause me irritation and headaches? Yes.
Did I want to waste my time on irritation and headaches when I could be doing other things, like learning the solo to ‘Time‘ by Pink Floyd? No.
So I politely told the instructor I wasn’t interested in his new technique. Other players might benefit, but I was fine with what I had, and it would be better if we just moved on to other things.
Trying To Get Your Hands Around It
That’s how I feel about 128 Technology. The company has some very good ideas about how to do routing differently (and again, you get all the myriad details in Ethan Banks’ excellent overview here).
Their approach makes sense, is technically feasible, and in some cases will provide advantages over traditional WAN routing.
For instance, sites that are bandwidth-constrained will benefit from 128T’s efficiency. Service providers and MSPs who want to be able to remotely provision and manage customers’ WAN routers will appreciate the Conductor’s capabilities around ZTP and configuration management.
The problem is that you have to go into great detail to understand how 128T does what they do, why they do it that way, and then parse all that information to decide if it’s better than what you’re doing now.
In other words, 128T asks potential customers to rethink what they know about routing, to change their mental framework, and get their hands around new concepts and techniques.
Certainly network engineers are smart and capable people who can learn new things. But they are also busy people with lots to do. They have to weigh costs and benefits of such an exercise.
If you have a clear and compelling need that matches 128T’s advantages, then the costs may be worth it. If you’re deeply frustrated with your current WAN routers and would love something different, take a look at 128T.
But if things are fine and packets are getting where they need to go, and you have other things you need to get done (like learning the solo to ‘Time’ by Pink Floyd), 128T’s uniqueness may not be worth the effort to change your chords.
