Last time, this column (Sept. 8) framed the context for Cablevision’s business-services arm, Optimum Lightpath, to begin hauling big video traffic — using “carrier Ethernet.”
This week’s translation aims to untangle the nomenclature oozing from the commercial-services sector — “carrier Ethernet,” for example.
Or try this glittering gem of techno-phrasing, which came up during the research: “It gets encoded into IP, from ASI or whatever, and then it gets tunneled through an MPLS LSP.”
Also rampant in carrier Ethernet speak: layers. “Dedicated layer 2 point-to-point,” for instance, appeared in the Optimum Lightpath release. We’ll get to that.
AREN’T ETHERNET & VIDEO AT ODDS?
Let’s back way up. As noted last time, “carrier Ethernet” is partly about sending digital video in a different way than it’s normally sent, for residential cable use.
Both methods share one attribute: They’re both about moving digital stuff (packets and bits) from one place to another.
It’s apples and oranges after that.
For example, Ethernet was born digital. Video evolved into digital. Digital video is data, yes, but it’s data with special needs, compared with other kinds of data, like e-mail, or the bits of Web surfing.
Ethernet began as a way to connect all the computers and printers in offices. Digital cable began as a way to send digital video to residences and consumer TVs.
Ethernet became “carrier Ethernet” when Ethernet itself surged beyond offices, into the larger realm of wide area networks — now the rings of fiber encircling towns and cities, owned by telecommunications carriers.
As a shared conveyor belt between office PCs, Ethernet began as a “collision-based” method. Sharing a wire means data might collide from time to time, so, early Ethernet had a built-in recovery mechanism. In collisions, each sender was to back off, stay quiet, and re-send.
That’s fine for e-mail or Web surfing. Not for video. Imagine it: Some bits in your stream are trying to get to you. They accidentally smash into other bits going elsewhere.
On (pre-carrier) Ethernet, your bits knew only to stop, twiddle their virtual thumbs for a few nanoseconds, then try, try again. On your TV screen, you’d see the video, and the words herky-jerky might come to mind.
That’s why quality-of-service (QoS) mechanisms are a big part of the carrier Ethernet picture, especially for sending digital video.
SPEAKING IN LAYERS
Remember that “dedicated layer 2 point-to-point” mention? In this sense, “dedicated” means that it uses QoS. It means “just for you, Big Video Company.”
“Layer 2” is a red flag for data speak. Specifically, data people speak in any of seven layers. (Still working on the bean-dip metaphor here.)
“Layer 1,” to an Ethernet person, is the “physical layer.” In cable, we’d call that “HFC” — hybrid fiber coax — the actual, put-your-hands-on-it plumbing of any contemporary cable system.
“Layer 2,” in data-speak, is “the data-link layer.” It’s how Ethernet networks define how to transfer stuff between nodes on a LAN — hence, “point to point” — but not how that data got to the node, from some bigger network.
Layer 2 is where the similarities dim between carrier-Ethernet transport and residential digital-video distribution. Why? Because digital cable is a one-to-many transmission, not point-to-point. Even though HFC architectures are designed around “500 home nodes,” those nodes aren’t equipped to send data between them.
That’s also why it’s so easy to feel dumb when data people use phrasing like this: “Are they really moving packets around at layer 3, or are they more acting like a layer 2 service and just moving frames?”
You didn’t miss a dot. In the residential-cable video world, there isn’t really a layer 2 or 3.
And what about that “MPLS LSP”? It stands for “Multi-Protocol Label Switching/Label Switch Path.” (Again, technical people tend to name things for what they do. Forgive them.)
MPLS LSP means that, if you’re a video packet running on a (carrier Ethernet) network that speaks the MPLS protocol, you first get tucked (“encapsulated”) into a tunnel, to keep you safe.
Next, you get a label stuck on your forehead that says where you’re going. Relative to “regular” Ethernet traffic, you’re in the express lane. The “label switch path” part is what looks at your intended journey, to figure out the most direct route.
So, translated, that crazy MPLS LSP sentence that started this discussion means this: Big video companies considering this new way (carrier Ethernet) to haul their products don’t have to pay to change any existing interfaces. The digital bits that constitute their video products get tucked into an express lane with a built-in nav system, working ahead on the fastest route.
Postscript to “Secret Bandwidth”: The mail and commentary continues on that Aug. 11 column about the “secret bandwidth” of addressable advertising. Several readers wrote to let me know that switched digital video implementations can tuck in addressable ads without using tangible 6-MHz channels. So, “secret bandwidth” applies if the splicing-in of the targeted ad happens in the set-top but not if it happens “higher up” in the network — in the switching fabric. Which raises the question: If an advertisement is addressed in a forest and no tangible bandwidth is used to carry it, is it still a secret?
Stumped by gibberish? Visit Leslie Ellis at www.translation-please.com.