Slashdot Mirror
The Risk of a Meltdown In the Cloud
zrbyte writes "A growing number of complexity theorists are beginning to recognize some potential problems with cloud computing. The growing consensus is that bizarre and unpredictable behavior often emerges in systems made up of 'networks of networks,' such as a business using the computational resources of a cloud provider. Bryan Ford at Yale University in New Haven says the full risks of the migration to the cloud have yet to be properly explored. He points out that complex systems can fail in many unexpected ways, and he outlines various simple scenarios in which a cloud could come unstuck."
At some point, there is going to be a massive failure. Someone big is going to lose *all* of their data. I still don't trust virtualization despite it being years old. It's still nascent in the grand scheme.
Someone wake me when they invent the holodeck.
I don't understand the intent of the article other than to provide a knee-jerk chicken-little response to cloud processing and storage.
Not one of the items mentioned is unique to the cloud. It can happen to any data center with more than two nodes involved in a cluster.
But that's not surprising, because "the cloud" is just a distributed collection of cluster servers, the same as large multi-nationals have been running pretty much since their customer loads exceeded the ability of one server to span the global community.
I do not fail; I succeed at finding out what does not work.
The TLDR version of the article is that load balancers can oscillate.
Its spun into a cloudy-thing because thats trendy, but the basic argument is nothing new.
Perhaps there's more "meat" in the original paper?
One common thread is that nothing is ever really "new" in computer science / IT. Clouds are just a rehash of ye olde mainframe outsourcing from decades ago. I worked at a place that was doing that in the early to mid 90s.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
I think the safest bet is to have local copies in addition to copies in the cloud, even if all the processing and computing is actually done in the cloud. Companies should set stuff up to keep a local copy of critical services on a good old fashioned tape drive or backup server. This is sort of a reverse of the cloud based backup solutions, where local processing and databases took place on local servers, but had backups in the cloud in case of a local disaster. Same idea: Have a local backup in case of the meltdown of the cloud. You may find your primary app is temporarily useless, but you at least have all your critical data (hopefully in a format that can be transferred.)
Occasionally living proof of the Ballmer peak.
Well, it's funny to me. But I was thinking that what you've got is a bunch of like computers operated by a single organization, connected by management networks, and depended upon by thousands of other organizations. If there was ever a sweeter target for a virus I don't know what it would be.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
What if someone trips on it?
I think the best metaphor for cloud computing is banking. There are risks for doing it yourself and risks in trusting a bank. If you hold your money in a safe then you may be robbed by a family member, employee, or petty theif. If you keep your money in the bank you may be robbed by the goverment, bankers, or by proffesional bank robbers. This same dynamic is true of storing data. The big difference between cloud computing and banking is regulation. Federally insured banks are heavilly regulated. There are laws about home much money they can lend, who they can hire, and what rates they can charge.
A coworker discovered when he upgraded from Kindle 1 to Kindle 2, many of the items he had purchased were no longer in the cloud (as amazon had promised). Most of what he lost was periodicals like magazines, but also some books. He was not a happy camper and asked for a refund for those books he could no longer acces, but Amazon simply told him they are not responsible.
That was back in 2009 if I recall correctly so maybe some of the bugs have been worked-out, but I stored it in memory as a reason why I won't trust the cloud to store any books I might purchase (or anything else). I try to back up these things to USB drive and googlemail storage.
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
Complex systems almost always exhibit surprising behavior. Cloud computing is no exception, and it is new in addition. This leads to a high level of risk of such events emerging without warning. Of course, people with a stake in the business side will never admit the risk. For examples of this happening in other fields, look at TEPCO, BP, RSA, ... All save and risk-free. Until things blow up.
Put simple: "The Cloud - where other peoples servers can crash yours."
Also appropriate:
"A distributed system is one in which I cannot get something done because a machine I've never heard of is down." --Leslie Lamport
This holds even more for the cloud.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
I've always thought that the problem would be when companies start using cloud services that rely on cloud servers that rely on cloud services and one of those cloud services has an outage.
A problem at one cloud provider can trickle up and affect your service that's running on a completely different cloud service, so for example, your website running on EC2 depends on order fulfillment software running on Rackspace's cloud, which uses back-end software on MS's Azure cloud.
If Azure has a hiccup, then your web store goes down, and if enough sites are affected, it can make real changes in the load (maybe less load because people can't shop, maybe more load because users keep hitting "reload" to try to place their order) on EC2 and Rackspace which could cause additional problems as load balancers try to shift load around as they respond to the sudden and huge change in load.
I do not know how broadband will be able to cope with this. No thanks, I try to keep independent and use good old rsync to keep my machines in sync. If things continue as they are we soon have bandwidth caps. Relying then on the cloud could become very expensive. Not only because of prize hikes in the cloud once the public is hooked, but also because faster internet service is needed.
I've been working in the cloud since July. The company I work for really likes the idea of it. But I'll tell you something. As a programmer and systems administrator responsible for something that lives in the cloud, I'm just not seeing the value of it. At least the way it's implemented at Rackspace. We've had problems that are absolutely bizarre, that seemingly have no explanation, that take weeks to resolve, that don't originate on our side. We've had issues with data integrity that don't happen on regular servers, and while we're able to "scale," we're very limited in the ways we're allowed to do it. Maybe this kind of set up works for other companies and groups, but I can't see myself choosing a cloud provider over traditional collocation and the standard three tier server model for 99% of what I need to do.
Just to add a footnote: it's interesting to note the differences for this search term using google vs. bing.
I'll see your senator, and I'll raise you two judges.
. . . until it does. No one with a vested stake can afford to speak the unspeakable of the greatest boon to the networking industry in the 21st century. This is the goose that is laying golden eggs, and the entire industry is cashing in on it. Before it is over, the bulk of all data will have shifted to the cloud at a cost of over a trillion dollars. The only people who can speak ill of this are those who have not learned to profit from it. Someday, there will be a great business in getting everyone out of the cloud. The marketing slogans can be easily predicted, "Get you head out the clouds and data out of the storm . . . don't let lightning strike . . . why is your precious corporate information floating away in the cloud?" Until then, everybody rides the train.
Of course, it may eventually fail in a spectacular way, but it will weather the occasional failure (see Amazon's). Then again, California will have a 9.x earthquake someday, too. However, you make notice that the land values correlate closely to the major fault zones, too. Ironic, but such is life.