Forget 'the Cloud'; 'the Fog' Is Tech's Future
online.wsj.com
online.wsj.com
(paraphrasing) 'In the rush to 'the cloud' people are disovering that there are bottlenecks related to bandwidth to edge devices (phones etc) and this problem will only increase as more and more devices become conencted (cf Interent of everything). However, not all services/products need to constantly push data back and forth to a centralised place. There is enough processing power in many devices such that more processing can happen at the edge so why not build more applications this way? Cisco and IBM see merit in this approach and hope to capture part of a (presumabley nacscent) market by providing some kind of processing ability that is a little nearer the edge of the network, rather than huge datacentres.'
Is that any better?
My view on the article is that until developers can more easily build distributed applications, then we're stuck with the prevailing paradigm of large centralised services. I don't see how anything from Cisco or IBM changes anything other than having a slew of mini-datacentres that essentially do the same thing.
What we really need are the means to create, deploy and maintain software in a more distributed manner. Having all the fancy smart sensors in my home talking to my hub is preferable to having them all talking to their individual silos. At least then, I can excersice some control of what's going on within my network and benefit from e.g. my high-speed wireless vs my much slower upstream bandwidth.
FWIW I'm working on an open source toolstack to make building distributed systems a lot more stratightforward. Essentially, we want to get to a place where everyone can own their own piece of the cloud. Whether that ends up being called the 'fog' or 'dust cloud' or 'motes' is beside the point. You can read an overview at http://nymote.org/blog/2013/introducing-nymote/
... and thus spins the endless wheel of IT. You could find this exact same sentence in any number of trade journals going back every five years to 1970. Just replace "mainframe" with "cloud" and "client server model" with "fog".
See you in another five years...
-first there were mainframes with glorified type-writes as terminals
-then we had personal computers
-then we had thin clients
-then we had thick clients
-etc
Basically this has happened before, will likely happen again, yet people (or maybe just tech reporters) always seem so surprised when the pendulum swings back the other way
The fact that the pendulum is swinging back the other way is news, though maybe not very exciting news.
I would hope for more applications that use cloud service to hook up apps on mobile devices, that would then directly peer to peer. I have fairly broad experience in networked applications since the 1980s, but this is one application area that I have never got my fingers on except for peer to peer UDP connections in a few networked games (LAN only). It will be really interesting to see the future effect of IPv6, more use of lossy but more efficient broadcasting protocols, etc.
If there's one thing you can bet on in computing, it's that technological trends are a pendulum that swings back and forth over time. The pendulum has been pegged on the "cloud" side for a long time.
One thing I've learned over the years is that the pendulum changes direction pretty much at the precise moment you start thinking it's not going to. I think "the cloud" falls in that category right now. I don't really know what the swing away from cloud computing looks like, but I won't at all be surprised when it shows up.
Then by your own logic, it's still a long way off.
That said, I don't know if there is a snazzy marketing term, but offline-capable apps with proper syncing support is the best solution for most stuff I want to create and use. I have a super computer in my pocket, so yeah it's nice to have snappy stuff done there, but data kept nice and backed up (as well as dispersed across the rest of my devices) using the "cloud".
Calling it "the Fog" is dumb, however.
> I'm as big a believer in the transformational power of cloud computing
> as anyone you'll meet. Smartphones, which are constantly seeking and
> retrieving data, don't make sense without the cloud, and any business
> that isn't racing to push its data and software into someone else's data
> center is, in my view, setting itself up for disruption by a competitor
> who is.
- My local restaurant wont get disrupted because it's not using Cloudfront.
At first I though it was a parody (Seriously, "the fog"?), well apparently I wasn't wrong...
To put it out there, I was in an unfamiliar city this weekend and when my friends and I wanted to order some food, simply for convenience: if they weren't on grubhub they were off our radar; so I'd say "the cloud" is even offering a competitive advantage for your local restaurants that use it vs those that don't.
In principle, I agree but that's not the way things are going. How many apps do you have that have to talk to (someone else's) backend service to be at all useful? For me I can think of Google Maps, Citymapper, Whatsapp and the weather app (just off the top of my head). All these are useless to me without a data connection.
To me it's especially the "don't make sense without the cloud" bit that I find egregious.
I'm trying to argue that we're heading toward a system where services that we use daily don't function well without some form of data connection, even if that's only for the purpose of synchronisation (e.g. your notes being available when you get to your tablet/laptop). I agree that smartphones can work perfectly well without the internet but that's not quite the way we use them now.
We might think we'd be fine offline but try turning the wifi and data off on your smartphone and see how frustrating you find it.
Where did you get music? Scrape from CD?
As for GPS and calls, their infrastructures are also part of the Cloud.
Cloud, fog, internet of things, big data, smart, edge computing. Ugh.
Noting to see here, just more PR fluff.
All joking aside, this seems like a weird term to try and coin. Isn't it just what people are mostly calling the "hybrid cloud" today?
It's about control of my data; I have to give it up to to benefit from a computing service.
Homomorphic encryption is an interesting thought: Allow my neighbors or a computing service to process my data without them able to see it. http://en.m.wikipedia.org/wiki/Homomorphic_encryption