CAD Cloudfusion (Part 2)

say-what-again-i-dare-youLast time on E(E) in Part 1 of Cloudfusion, we outlined how the ongoing cloud bombardment is certainly creating plenty of disruption and confusion in the Computer Aided Design (CAD) market.  If you missed Part 1, catch yourself up right here.  Tired of the word cloud already?  You’ll have to endure a few more invocations, I’m afraid.  In an attempt to diffuse the ongoing cloudfusion, we’ll dive right into defining the various cloud technologies that are relevant for CAD.  Here we go.  In no particular order:

Cloud Software Delivery:  Pretty simple actually.  Instead of receiving your software on a piece of media, delivered by a mule along the Inca trail, or air-dropped by one of those Amazon Death Bots quad copters, the software is delivered and updated via an internet connection.  That doesn’t require the software to maintain continuous connectivity nor does it force files to live online.

Cloud Applications: Otherwise known as Software as a Service (SaaS) or more commonly as the web app.  The software itself lives in the cloud and typically so does the data, but that doesn’t necessarily mean it must be over the intranet.  In the appliance model (no not like your toaster), the SaaS can live in an on-premise black box on a private network.  Could an SaaS work in an offline mode as well with locally stored files?  It sure can, as long as the developers support it.  Google Docs can, as long as you use Chrome (see what they did there?).

Cloud Collaboration / Sharing / Storage:  A few technical types will furrow their brow at this point and note that cloud sharing and storage like Dropbox really is just a particular example of SaaS.  I’ve put this in a different category here because A. The SaaS functionality is really just to provide the gateway to cloud storage and B. this class of SaaS is more likely to work transparently with conventional desktop applications.  This variation of cloud is probably what most people are keenly aware of, because in many ways it’s the least disruptive.  Cloud storage might be synchronized with conventional storage or might require manual intervention.

Cloud Platforms: If cloud was an underground bunker, think of Platform as a Service (Paas) as going one level below SaaS.  So instead of just offloading an application into the cloud, PaaS offloads a whole computing environment, primarily to foster the development of SaaS applications.  So you can stand up your really fat database without actually having anything physical to verbally abuse or kick when it doesn’t work.

Cloud Infrastructure: Infrastructure as a Service (IaaS) is the lowest level in the cloud bunker.  This is offloading at the hardware level including storage, networking, etc either as physical or virtual machines.  This level is for really for just aggregating raw computing power without buying or maintaining it yourself.  That buys you the flexibility to store your own virtual Johnny Depp, or perhaps your own Lawnmower Man depending on your preference of singularity monster.

Cloud Processing:  Elastic computing is the official term, but it has nothing to do with rubber bands.  Elastic computing allows hardware resources to be allocated on the fly to meet computational demand.  Sometimes this is referred to as the elastic cloud.  Obvious applications are in the CAE space or in large model rendering.  Recent technology has demonstrated concepts where local computing and elastic computing resources can work in concert within the same application, whether that application is a local application, an SaaS or a hybrid of the two.  Confused yet?

Virtualization: Ah virtualization, where most people’s heads explode.  Virtualization is turning physical computing assets and software into virtual ones.  Virtualization allows any of the above to be stuffed as a file on any capable hardware or inside another virtual environment.  Virtualization can be really useful for security and enabling elastic computing, because all the hardware or software requirements can be wholly abstracted.  That means you can operate what would otherwise require the mother of all supercomputers on what’s essentially a dumb terminal with no storage – or your mobile device.  The major drawback of virtualization: virtualization into the cloud introduces considerable latency.  While creative solutions address these limitations, you still need a really fat pipe to ensure smooth operation.  Incidentally virtualization helps foster emulation but the two are not interchangeable terms.

You can see that all these cloud technology variations aren’t necessarily dependent nor mutually exclusive.  Could you have a web app CAD program that stores files locally?  Could a desktop application utilize elastic computing to offset high computational loads?  Might a cloud CAD collaboration platform exist as an appliance inside a company’s on-premise server room?  Could you virtualize all of the above into an entirely different private cloud located exclusively in suburban Milwaukee?  You bet your bananas on all counts.

So the important lesson here is avoid looking at cloud technology as an absolute feature set or a specific implementation thereof.  Sometimes this is hard to do while in the midst of the marketing onslaught.  Cloud involves a wide repertoire of technologies.  While commonly bundled together, there are many diverse ways to slice the cloud pie.  Mmm… cloud pie.  Expect the variety and choice of cloud technology to continue to accelerate from here.  Have more cloud conundrums?  Drop them in the comments.  Good luck and cloud on.