After Cloud Computing comes Edge Computing

Can you say, "Pivot," again?

 

Software companies like Autodesk, Graebert, and Onshape have gone all-in for the cloud. Others, such as Siemens PLM Software, Bricsys, and PTC, are cautious about jumping on a bandwagon. Dassault Systemes is a straddler, with V6 being cloud, but V5, Solidworks, and DraftSight firmly on the desktop. 

The development was logical: Computing moving from standalone PCs to proprietary local area networks (office-wide networking), to standardized LANs (twisted-pair ethernet), to Internet (worldwide networking), to WiFi (untethered networking), and finally to cloud (multiple servers online).

But design offices are not homogeneous, as much as some marketing departments wish it were so. In fact, PTC stated it bluntly in a recent conference call: only new companies and large companies are interested in the cloud; everyone else isn't.

On the other side of the equation, CAD vendors aren't necessarily able to deliver a cloud-based program. While Onshape has a full-ish MCAD system working in the Web browser, it started from scratch with the assistance of $168 million in outside funding. 

The more common experience, however, is that legacy firms are hitting barriers, as much of the desktop code is locked to Windows APIs. A decade after it launched V6, Dassault says only 2/3 of its V6 software is ready for the cloud. Autodesk is also well behind, failing its former CEO's goal to be all-cloud by 2015 or 2016.

Maybe hesitation and delay are good things. Because now comes news of the next advance in computing, edge computing. "Edge" means computing at the edge of the network, such as  in drones, robots, cars, and the "sheer volume of the Internet of Things." General partner Peeter Levine of venture capital firm Andreessen Horowitz warns:

Cloud computing is soon going to take a back seat to edge computing, and we will very quickly see the majority of processing taking place at the device level.

To understand why, consider how data is moved around inside your computer between the hard drive, memory, CPU, and graphics board -- very quickly over high-speed buses with pretty much perfect reliability. (Most crashes you experience with your computer are not the the fault of the computer, but due to the software running on it.)

Now think of the Internet as a data bus between your computer and remote servers -- a slow, unreliable bus. My hat goes off to companies like Onshape and Graebert that work very, very hard to make it seem like the Internet's fast and reliable.

As Mr Levine points out, computing moved from server-client once before, albeit slowly (in the 1980s). With the quickened pace of technology, the second move from server-client (cloud) will happen in less time. "If a car needs to a make decision, it needs the information instantly and no amount of latency is going to be acceptable."

You may argue that designing a widget is not a task that needs edge computing. Actually, it does. The graphics board in your computer is a high-speed powerhouse that makes real-time rendered 3D motion possible. It's an example of simple edge computing. Preliminary work has been undertaken for years now to make that GPU perform more computing, except that it works best on parallel operations, of which CAD has few.

Nevertheless, CAD vendors can take heed of Mr Levine's warning that the cloud is not the be-all end-all utopia that some executives insist it is.

With data breaches operated even by governments formerly considered trustworthy, I can already hear the marketing line pivoting in a few years from now: the cloud is slow and insecure; only legacy companies still trust their data there.

- - -

Source: https://techcrunch.com/2017/08/03/edge-computing-could-push-the-cloud-to-the-fringe/

 

 

 

Comments

[Tom Foster]

Is Edge Computing the final coming-of-age of the 'democratic' Distribted dream of computing, inherently inter-operable and bypassing centralised gatekeeper(s)?
If so, why is a Venture Capitalist like Peeter Levine so interested, without plotting yet another ingenious way to create ever-stronger super-rich monopolies?
It was dismaying that the Social Media giants so cleverly trapped inherently (we thought) inter-connected internet users into monopolistic siloed traps, to data-mine and re-sell their lives, as described in Ruskoff D - 'Throwing Rocks at the Google Bus'.
Will empire-building entrepreneurs and their Venture-Capital drivers create out of Edge Computing yet another even stronger form of dominance over 'us'?
Or is this really part of the end-game of Capitalism, as hyper-productivity dries up the profit motive and a Distributed Commons-based economy grows, as foretold in Rifkin R - 'The Zero Marginal Cost Society'?

[Ralph Grabowski]

When you read Peter Theil's 'Zero to One' you learn that the plan always is to turn a service into a monopoly. The 'zero' refers to no business and the 'one' to the business as monopoly, like PayPal.

[Tom Foster]

Yes indeed but there's a hope around that that perennial phenomenon is a function of profit-motive capitalism, and if profit as a driver is drying up - Jeremy Rifkind's (and others') future-gazing perception - then 'winner takes all' as biz model may no longer be possible and populations will no longer have reason to collude in that.
Or will new forms of monopoly be invented and will 'we' allow it?

[Nerd 4 Hire]

As an autodidact who fixes computers using self-trained methods, I've never come across a proper cloud verses edge computing before. Thanks for taking the time to do this as I learned a lot!

Sincerely, Wheeler from Computer Repair Nerds

[Robert James]

That’s cool that managed cloud service providers can help simplify operations. The greatest things in life are simple. I've never come across a proper cloud versus edge computing before. Thanks for this.