The Joel on Software Discussion Group (CLOSED)

A place to discuss Joel on Software. Now closed.

This community works best when people use their real names. Please register for a free account.

Other Groups:
Joel on Software
Business of Software
Design of Software (CLOSED)
.NET Questions (CLOSED)
Fog Creek Copilot

The Old Forum

Your hosts:
Albert D. Kallal
Li-Fan Chen
Stephen Jones

What I Don't Like About Cloud Computing, Episode 6

Bruce Schneier should be familiar to anyone who is concerned with IT security.  This video,297151,sid14_gci1355568,00.html is a point/counterpoint discussion with Marcus Ranum, who also has a pretty good pedigree in IT security.

There are several things I took away from this, certainly vendor lock-in has to be at the front of the line for reasons not to use cloud computing.  Another point is reliability, and the suggestion that the reason Google's support for things like GMail is so poor is because you are getting what you pay for (i.e. nothing).  Schneier's first point is probably his best, that cloud computing is nothing new, only that the network has moved onto the World Wide Web.  That means that cloud computing is going to stir up the industry for a few years, but it is hardly a revolution.

I'll add my two cents worth, and make the point that all of the benefits of cloud computing accrue to the provider, and users end up with all of the disadvantages.  The user gets slower applications with no control over his or her own data, has to rely on the trustworthiness of the provider, and because it is a painful process to switch to another provider, has very little leverage to obtain better service with.  Bandwidth still comes at a price; what desktop computers can't compete with the cloud in terms of processing power and storage at the level of individual users?  Shouldn't the software industry try to improve the user experience, not degrade it?
RGlasel Send private email
Monday, May 25, 2009
"...and make the point that all of the benefits of cloud computing accrue to the provider, and users end up with all of the disadvantages."

I generally agree with RGlasel's thinking.

I wonder if there ever will be a killer app where the advantage/disadvantage ratio will be much more balanced and fair?

I suspect that the folks pushing cloud computing are really pushing the infrastructure, and are simply relying on programs like Google Docs and Office Online to stress test the infrastructure.  They want to in effect, become the new telecoms.
TheDavid Send private email
Monday, May 25, 2009
i'm sorry if i don't get it all, but i thought cloud computing was all about infrastructure and all these servers that you can't buy?

then for me it's a win.
Victor the Python Artist Send private email
Monday, May 25, 2009
I agree with critics. The huge economic advantage of renting an application/infrastructure instead of buying it quickly fades when it comes to things like vendor lock-in, control over data, tricky SLAs etc.

It is nothing new but for a few specific infrastructure (Google, Amazopn and Microsoft) that now are a development option. None the less I'd never rely on them for my mission critical data.
Stray__Cat Send private email
Monday, May 25, 2009
My perspective is that cloud computing is moving processing and storage from the user's computer to a remote server at a different domain.  I suppose cloud computing could be a win for users (who may or may not be developers) if they could only afford thin clients for their own computers.  These days there are kids in Nigeria melting down motherboards that could provide more processing power than the slice of a server you get when you use GoogleDocs.

When I tried my hand at a couple of CS courses in '80-81, there were about 250 terminals on campus connected to about a dozen PDP-11 mini-computers which communicated with a single Amdahl V8 mainframe, and the network ran on the Michigan Terminal System.  Most of the 250 terminals were essentially teletype machines, you could only edit a single line at a time, and once you hit Return, whatever you typed was gone forever, unless you sent a command to have a previously entered line printed on your terminal.  You would end up writing as much on your (paper) notepad as you typed on the keyboard.  There would be lineups to use the 30 "smart" terminals that could store 20 lines of 80 character text in a screen buffer.

I don't know how many people realize what an incredible advance personal computers are.  Unfortunately, software development hasn't been able to take advantage of modern PC's, otherwise, why would anyone waste their time with dumbed down word processors hosted remotely and displayed in a web browser?  I do think there is a place for shared remote storage and relay agents that can connect computers almost anywhere in the world, but I just don't get the movement to cloud computing (unless someone thinks they can get rich at the expense of all those dumb users).
RGlasel Send private email
Monday, May 25, 2009
This really sounds more like Software as a Service than cloud computing, if you're complaining about locking up data in things like Google Docs.

And even then, I very much like Google Docs for those things that I might need *anywhere*.  I use it for documents only I will be using, to say nothing of documents that are shared.

I'm not sure how vendor lock in is any worse with cloud computing.  It's far more frequent that there are export facilities or access to a full blown SQL database for data export than with any number of other apps.
Lance Hampton Send private email
Tuesday, May 26, 2009
Yep, agree about cloud computing, it's got severe drawbacks and no real upside for 99% of projects.

This is just the 2008-09 flash in the pan vendor promoted meme. Nothing new. Just ignore all this stuff until it's been around at least 5 years and you'll save a lot of time since most of these take at least 3 years before the totally obvious problems that you got ridiculed for pointing out become accepted known problems.

Look at all the idiots that learned Java, or the dumb schools that converted their curriculum to be Java trade schools. Now Sun is dead and Java has still not gotten its act together. What now.
Scott Send private email
Tuesday, May 26, 2009
In anticipation of the flames I just baited, consider Perl. Perl is a lot less complex than Java. Where is the next version of Perl? Well it is out having drinks with Duke Nukem Forever. In other words, it is dead. In the case of Perl, the current version is fast and works. The next version will be slow and even have different syntax, so it's a totally different language. Come on now, even if delivered, you can see where this all is headed. CPU speeds are not doubling any more, in fact they are not improving at all and haven't been for years. That extra performance needed isn't coming and yet the language and library designers think they can just keep adding crap and bogging stuff down. No more, that game is over. Java never got to the point where it worked properly, so now that it has no advocate to direct it and fix the problems, it will die. Slowly, like Cobol, but within 2 years it will be common knowledge that it is pointless to start new projects in Java. And we'll have all these college graduates that  don't understand pointers and can't write in C or C++ or even Perl.
Scott Send private email
Tuesday, May 26, 2009

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
Powered by FogBugz