What can you say about the edge? It’s a new frontier for digital infrastructure. Everyone is talking about it. And everything you hear is almost certainly exaggerated.
The big pitch for the edge is as follows: applications are increasingly distributed, in particular those related to the Internet of Things and streaming media consumption. Meanwhile, resources have been centralized in the cloud, which is delivered from large, remote facilities. When the distributed applications need a fast response time, they need local resources close to the users (streaming content) or devices (IoT). This means resources at the “edge” of the network.
I have no news but I must tweet
Beyond this pitch, there really is not much to say about the edge at this moment. The details are up for discussion, and the actual delivery methods are still being determined.
Except, of course, for one well-established category of service providers who have been doing it for more than 20 years. Content delivery networks (CDNs) such as Akamai and Cloudflare began caching Internet content on proxy servers in the 1990s, so users could read web pages more quickly.
Today’s edge network proposals have to do the same thing for an increasingly mobile generation of users and applications, and it is vital for any digital marketer to say something. Whatever your company does in the digital infrastructure, you must convince your customers and prospects that you are (if it’s not an oxymoron) absolutely central to this thing called the edge.
That’s why we’re seeing a myriad conference presentations and strands on the subject. That’s why we’ve been noticing hastily-compiled lists of Edge influencers, stuffed with people who may be surprised to find themselves labelled as such.
And of course, we ourselves have been adding oil to the fire, discussing what edge might be, who will be providing it, and what difference it will make. We have to: it’s our job.
Our correspondent David Chernicoff kicked off our coverage in 2015. DCD then put together an edge summit in Dallas in 2017 You can read a long article here about that highly-productive event, but here are some nuggets:
- Edge resources will result in sales of micro data centers (though more radical distributed resources should win in the long term)
- Although location is important, the key factor is response time. If it can live with several milliseconds of latency, your edge application could just as easily live entirely in the cloud.
- Resources at the edge are deployed in small, distributed elements. Without the economies of scale, they will be relatively more costly - in terms of power and money - so they will be deployed reluctantly
- The edge will therefore not replace the cloud in any sense at all. It will augment and support it. Those paying for the hardware will keep it to an absolute minimum
One really obvious location for edge resources is the “last mile”, usually meaning the network of cell towers which are used to deliver mobile network services, but also including the Central Offices (telephone exchanges) of the fixed phone networs. These resources already have power and networks on tap, so let’s just add some IT, in the form of a containerized data center, or some smaller box of servers.
Vapor IO and others have presented this as an opportunity for a kind of distributed colocation service, where servers are located in micro facilities at the base of cell towers. That’s a clear concept, and we could even predict the anchor tenants for such as distributed facility: Netflix, Amazon Prime and anyone else wanting to pump TV episodes to mobile devices.
As far as we know, mobile networks haven’t yet made a big move in this direction, but all the vendors have to keep themselves front-of-mind. Hence a lot of material promoting countless, fairly similar ideas.
Amongst this, the State of the Edge report, from a consortium including Vapor IO, Packet, Ericsson’s UDN, Rafay Systems and Arm, is a good effort to explain all this. Its 80 page report is worth a read. It is available here, and bits of it are also in the GitHub open source repository so you, too can contribute.
The report reckons that “the edge is a location, not a thing”. I’d say it’s more about the applications. It focuses on the last mile network, also known as the telco edge, of course, and makes one very good point: this is the interface between the network on one hand, and the customers and devices on the other. Edge resources could be on either side of this divide. Simplistically: in phones or at the cell-tower.
Where will these resources be? As with most dichotomies, the answer will probably be “both”.
After all that is said, we’re left wondering when this edge will arrive. If the need is that obvious, why isn’t it here now? The answer may have something to do with the uncertainties we touched on. Where exactly will it be, who will charge rent, and who will pay for it?
How much of edge will be in our pockets and how much will it take from our pockets? Only the gradual process of commericialization will decide this.
[Apologies to anyone reading this who expected it would have anything to do with the late Harlan Ellison’. I’m sorry to say the title is just a lame, subliminal nod to his best-known work]