Tuesday, May 26

Computing on the Edge

I attended Edge Congress last week in Austin, and the conversations were focused on computing on the edge. This post outlines what I learned at the event. In full transparency, I was given a press pass. However, no one has reviewed or approved any of my posts about the event.

Edge is an important part of digital transformation, and traditional ops people should be paying attention. This is one of the areas that is going to change computing, and you don’t want to be left in the land of legacy.

Computing on the Edge – you already have the tools

The host of the event encouraged everyone to think about three elements of edge:

  • For Developers: what networks does your app need? Have you made sure the app is able to transverse all the networks that you (or your application owner) expect?
  • For Developers/Application Owners: Define Latency. What does real time really mean to you?
  • For Developers/Application Owners: Define the importance of your application – what happens if something goes wrong?

For operations folks – doesn’t this sound familiar? This is where we try to get with current datacenters. Container-based applications, especially if deployed to on-premises and Edge locations, will make this exercise even more important and definitely more complicated. Data center hygiene will just become that much more important to the Edge.

Don’t rinse and repeat at the Edge

The first speaker was Joe Reele of Schneider Electric. His advice: we have to change the way we’re designing and deploying for the edge to reap the benefits. Said another way, we can’t look at Edge as an extension of the datacenter, and we have to start looking at Edge with the same economies that we have learned to apply to the cloud.

He explained the paradigm shift from old infrastructure deployments to self-reporting hardware that is managed as a complete microsite. His description sounded so much like VMware Cloud on Dell EMC that I had to ping Kit Colbert.

https://twitter.com/gminks/status/1192100934542860288

He also encouraged us – operations – to build out the capability on the Edge Don’t look for use cases, they haven’t manifested yet, and won’t until developers and application owners understand what’s being built out for them.

We need to build Edge as a total package that doesn’t increase risk to the application owners if we want to reap the benefits of innovation that we’re being promised. If we build it out like the same old edge and remote locations, there’s really no compelling reason to use it.

https://twitter.com/gminks/status/1192102030082486274

Where *is* the Edge?

This question came up several times. To many giggles and laughs. Of course, this is a question operations will ask, we have to build it and nail it to the ground after all. But that question was compared to the “where is cloud” question of a decade ago.

Edge can be where ever applications need the benefits of low latency, or bandwidth. There was even talk of re-purposing older datacenters to become Edge locations.

Don’t fixate on the physical where. Don’t get trapped into thinking this is the same as CO-LO or ROBO. Think bigger, think differently.

Edge use case: Fix gaming lag

One of the great use cases was from Mathieu Duperre, founder of Edgegap. Edgegap is a B2B company that helps gaming studios solve one of the most dreaded problems in gaming – lag.

If you play on online game, lag is when the latency between you and the online platform is so bad that you tell the game to do something, but by the time it gets to the game server it’s too late, you lost!

As Mathieu explained, gaming is an almost $150M industry.

So beyond subscriptions, gaming has competitions for huge prize money. The last thing a gaming studio wants is for a famous gamer to play their game and experience lag while they are live streaming.

Edgegap uses all sorts of telemetry to make sure everyone playing the game – no matter what device they’re using to play, and no matter where the players in the game are located – have a kickass game.

One example Mathieu gave was using telemetry to recognize that at 2PM around a certain location, load on game servers would spike. This could be because all the kids were getting out of school and logging on. With this info, Edgegap can deploy game instances to a location closer to the kids, and they all could play lag-free.

Computing on the Edge – what Ops needs to know

Computing on the Edge is going to get a lot of marketing attention. But what is the reality for operations folks? Edge is a huge opportunity for us – this is our sweet spot. You’ll need to learn how to talk to your development teams, and to application owners (you should already be doing this).

If you understand what these teams expect, and what they want to build, you can build the environment that they need. Start there, and then demand that your vendors help you achieve that goal.

We’re going into a new world, but it really isn’t uncharted territory. We’ve been here before, but don’t get pushed in a corner. It’s our time to shine!

Please follow and like us:
error0

Leave a Reply

Your email address will not be published. Required fields are marked *