This page has been archived and commenting is disabled.
Amazon Cloud Crashes, Takes Part Of Web Offline
Curious why there are those with an online business, who believe that handing over their entire back office infrastructure to one company, aka "going cloud" may not be the wisest of ideas. Just ask all those websites that use Amazon's cloud service today, who suddenly went dark when the Amazon cloud crashed.
Amazon’s data centers in Northern Virginia crashed Monday afternoon, taking with it a number of popular Web sites, from Someecards, the quirky e-card company, to mobile applications like Flipboard and Foursquare.
Amazon reported having problems with the data centers in Northern Virginia. Those problems appear to have had a ripple effect across the Internet with several sites hosted on Amazon’s popular EC2 cloud hosting service also reporting problems.
Several frustrated customers took to Twitter Monday to complain that they could not get access to Web sites including Foursquare, turntable.fm and Flipboard. It appears that some of the affected services then affected services that, in turn, ran on them. Because they are all hosted on Amazon’s cloud service, there is a ripple effect. They all go down when the original hosting servers go down.
Last June, an electrical storm caused problems at the same Northern Virginia data centers and took down sites including Netflix, Pinterest and Instagram for a weekend.
“Like many other services, we’ve been taken down by the outage,” said Erin Gleason, a spokeswoman for Foursquare, the mobile check-in service. “Both the site and the app are inaccessible right now.”
Ms. Gleason said the company was still awaiting guidance and updates from Amazon about when its service might be restored.
“Hoping to get things back up and running ASAP,” she said.
And just like HFT has taken the entire market hostage, in the most insidiuous Stockholm Syndrome symbiotic relationship, so the cloud will do that same to the internet, in some cases with insidious crossover: recall that various stock exchanges are already using the cloud to host their services. Basically what will end up happening is that one day the mega-uber-cloud will crash, and the entire internet will be down.
Hopefully not on demand.
»
- 17943 reads
- Printer-friendly version
- Send to friend
- advertisements -
HAARP loves clouds!
Just wait for the Russia-sponsored EMP party, it's gonna be a blast...
But I was told that clouds don't exist in any one place, so it's physically impossible for them to crash.
I've even got it written on a shiny postcard!
And just where were the Iranians?
Coming to a bank server near you!
I've been warning about this kind of thing from the moment I first heard the "cloud" buzzword - I was like, you have to be kidding me, this is the dumbest Internet marketing fad yet.
But Reggie Middleton loves the cloud!
EH! All you FUCKIN HIPPIES! Stop hating on your machine overloads ;D
For serial though. HFT facilities are well protected. Quite literally, keeping their servers well protected from emps, power outages, hacking, and server glitches, is financially demanded. If something were to go wrong... They could experience a "knight capital" type moment. It's literally one of their top priorities. Thus, the "free market", keeps roaring on. LOLOLOL...
On another note, should an extremely large high scale X-class flare pop outta the sun... Well there's warning time, but the bots might have to be shut down hours before impact, or risk getting fried. Hell, I imagine one day the earth could experience a solar phenomenon so large that trading is actually halted for a day until the CME passes.
The 1859 Carrington Event caused major problems with the telegraphs for a day or two. The same event happening today would cause quite a few problems that will not be easily fixed.
God forbid I might have to write a reciept by hand, or read a book for entertainment.
You'll be able to read by the glow of the melting nuclear plants.
Dear Cloud Subscriber, Thank you for your request to recover damages caused by our negligence but if you have your attorneys check section 32.34.1.45 of the user agreement, they will see clearly that you have released us from any such damages past, present or future. Thank you and have a nice day.
If only it would come the bank that holds my mortgage....
Jubilee! :-)
@ NotA + 1
That's why I keep my database on two computers and five flash drives... Backed up, bitchez!
Keep at least one of your flash drives in an antistatic bag.
This explains why Bitcoin will bust, credit card use will tank, and getting a ZH fix will cause me to pull out the few hairs I have left. Keep cash on hand! When the electronic payment methods crash I can still buy my necessities by the pallet-load at Sam's for the hide-away.
+1 RR
when the digital money system crashes, you can have all the cash in the world....but good luck getting to Sam's and purchasing stuff. i'm pretty sure that Sam's doesn't use cash to purchase stuff to fill their wharehouses. better do it now.
Microsoft and others have been trying to push "the cloud" for a very long time. My personal opinion is that it's an attempt to turn the whole internet (including content) into a subscription service and "the cloud" is a buzz term for managers not in the know. IPV6 will be the real game-changer and not "the cloud".
The "cloud" used to be called a MAINFRAME and iCrap used to be called DUMB TERMINALS.
the rest is just marketing...
A cogent analysis. In the Old Days, there was a big mainframe in the basement. Users interfaced with the mainframe using dumb terminals ("green screens"). Then Microsoft, Sun, and the Internet came along and convinced everyone that "the network is the computer". All the processing power was pushed out to the desktop. Businesses spent billions of dollars putting PCs on every desk and filling rooms with servers. Head fake! Now the IT industry has convinced business they need to get rid of the PC desktop and replace their servers with large, powerful hosts running dozens of virtual servers (exactly like the, um, big mainframes running LPARs back in the day) and give the users "thin clients". The old model has merely been repackaged as "the Cloud". After business is done spending billions of dollars virtualizing everything or sticking it in the cloud, IT will figure out a way to convince them they need to pull it all back, and we'll start over again. The endless systole and diastole of the IT industry's perpetual scam.
really? you know how much REDUNDANCY is built into internet based communication systems? THE THING WAS DESIGNED TO WITHSTAND A NUCLEAR ATTACK you bozo. Sure...on the FOR PROFIT side this is kick in the nads. But the internet itself is actually TOO robust. Just ask President What's His Name from Egypt. That guy got Mega-downloaded man.
Steve does not like it. That is all you need to know!
During an on-stage discussion with Apple (AAPL) critic Mike Daisey, the Cupertino-based company’s co-founder Steve Wozniak expressed major concerns about the future of computing and the shift to the cloud, Agence France-Presse reported. “I really worry about everything going to the cloud. I think it’s going to be horrendous. I think there are going to be a lot of horrible problems in the next five years,” Wozniak said, adding that “with the cloud, you don’t own anything. You already signed it away” through the terms of service all cloud computing users must agree to.
Yeah, and just wait for the day that all your cloudia shit is unplugged, and you get a notice from your provider that for the sum of $XXxx, they will be happy to turn it back on. Payable in gold.
The Internet was designed to withstand a nuclear attack ... but it was never designed to guarantee that the subnetwork were your company's main trading server is in will remain connected to it or even up, nor will it guarantee that even if it does stay connected the speed with which it can be accessed is acceptable.
More in general, we've been busy undermining that high-resilience architecture for the last 2 decades, be it through things like main-trunks and level 1 providers (there are a couple of submarine cables out there that if they go, some places will be disconnected or back to dialup speeds), through application level architectures which are far less resilient (cloud computing being one of those, since it centralises a decentralised system) or purelly through ever increasing amounts of data traffic which have made a couple of routes and nodes mission-critical (sure, they can go down - as long as you're willing to go back to no video and taking 4 minutes to load a web page). Even something as old (by now) as the concept of a single server with a web address is not redundant without some special trickery (with hardware load-balancers and DNS servers, if you're curious).
Proper virtualization, ala containers is very much a pullback to the IBM360 days, only now you can put a whole OS and application set inside that timeslot.
But virtualization alone is not "the cloud" because "the cloud" doesn't exist outside of the minds of idiot marketing hypetypes and their sucker clientelle (management).
"But virtualization alone is not "the cloud" because "the cloud" doesn't exist outside of the minds of idiot marketing hypetypes and their sucker clientelle (management)."
Welcome to the 'Brochure Management' era where middle managers get spooked on the first page, and then buy whatever the happy faces are next to on the second.. and then get mad when their engineers smirk at them for being a total noob
Cloud + Govt = Easy spying on your own citizenry
Imagine:
Automated personality profiling based on music library
Automated warrants based on facial recognition against your personal photos
Nailed it...
Who needs a fucking warrant or anything else to get at your shit when it already is already stored and maintained on the equipment of a third party who is more than happy to hand it over for a few dollars.
Try and find a way to sync your phone and your Outlook (for example) via a direct connection (i.e. a cable).
No longer possible -- why? Because someone(s) benefit greatly when your entire address book has to go through their servers on its way back and forth between your phone and desktop and whatever other devices you have.
Sure it is more convenient this way, but you really don't have too much choice.
If energy gets really scarce (real or imposed) propaganda about the cloud will get even louder. Wait for it...
LOL! Russia would never do anything to harm the US. (Sac)
the housewives of DC went shopping all at once; so where's Penetta telling us we need to guard against extreme online consumerism!?!?! The horror
Must be the Iranian cyber terrorists...... lets go to war...
NO CLOUDS = BULLISH SOLAR PANELS!!
If you don't hold it physically, you don't have it. Right, ZHers?
Nope! I'm not holding "it" right now, and yes, physically it's still down there.
/sarc
So concise...Bravo +10
How many nines of uptime does this put them at now? 3?
"Sir have you checked your SLA on the service definitions....no guarentees and no responsibility on data loss (section 16.5)...drive failure responsibility of the client...(section 23.3)"
There be a shit storm brewin.
Either they haven't tested their failover or there is no backup plan. I'll go out on a limb and choose #B.
Even tried restoring 1/4 of the web from tape back up?
Elbow gets tired.
Can someone please shoot me, I see the Big O's wife every time I load a page.
pods
Do Not Track Plus... It's free and it's great..... Here
I have to admit, I do miss the "Meh" girl ad
Maybe she's having the big "O"
The Story of "O"
Not a pretty picture, Chewbacca shiverin' and shakin'...
To the Cloud! - Taco MacArthur
I would love for the internet to come crashing to the ground.
.
Ironically, then we would become a 'social network' again.
Yeah, like the good ole days when we only had candlelight after dark. When a gust of wind came thru and blew out our candles we'd all go out on the front porch to check on the neighbors to see if their candles had blown out too.
That is awesome. lol
I kept fireflies in a jar for Disaster Recovery purposes...
Thanks ZH, for revealing scam after scam after scam. Again and again and again. Unbelievable.
Ancient Greeze, Bitchez
Long pens, pencils, paper,and paperback books.
That's why this company deserves a 300 p/e.
Is that the reason? And all this time I thought it was their 1.5% operating margin.
Why don't they copy the data to a second location ...as a solution on the side. Must be a cost thing...?
because is just a cloud, it is not evan on the first location
Cloud NEIN.
Black hawk ; I mean Reddit down
Now if it was down for the count that would be good news, alas it will be back though.
Having all of your data centers in place defeats the purpose of a cloud. fucking morons.
DOH!
Non-redundacy FAIL.
Ironic how lightning took out 'The Cloud'. Great naming since clouds are full of water and static electricity, just what you want near sensitive electronics, right? I'd rather put all my money in 'The Goldman Sacks'...When it rains, it pours... :)
PS. I'm amazed that these companies have all their eggs in one basket. Nissan had redundant centers on both sides of the country almost 20 years ago, capable of being fully redundant. Even some of my smallest clients can remain online if their external host goes down...
It;s those "damn Iranians" again....Leiberman said so! LOL
The fact that their redundancy can't handle a single site going down is clownish at best. And users have to manualy restart instances? That shit should be automatic and fairly seamless. Someone was being cheap. Clusters aren't cheap for a reason.
https://aws.amazon.com/support/
Cluster fucks are free of charge, any time. This was people being cheap ratfuckers.
Buy vinyl!
Cloud Computing ... welcome to the rentier society SUCKERS!!!
But seriously... are we really going to go back to the mainframe terminal model..... with the exception that someoneowns the mainframe...
long WOPR w/cheese
It was a great idea, for revenue. It is just like CDS and MBS, nothing in the cloud, nothing in the pool. But they charged everyone 10x (100x) for 1 promise. Hard to imagine why anyone wants their business in a cloud? All wet and you can't see and it is way the hell up in the air. Only Baumgartner would have a nice time.
Hey! You! Get off of my cloud!
I was sick and tired, fed up with this and decided to take a drive downtown
It was so very quiet and peaceful there was nobody, not a soul around
http://www.youtube.com/watch?v=jgWUi-ozMAU
Amazon's customer service dept appears to have their heads in the cloud as well. Cue Rolling Stones "Satisfaction".
http://consumerist.com/2012/10/22/amazon-erases-customers-kindle-wishes-her-luck-in-finding-somewhere-else-to-shop/
Flipboard and Foursquare?
Who gives a shit?! Reddit is down!!!!!!!!!!!!!!!!!!!! :(
Because the power congestion in Northern VA is getting out of control.
The Crashing Clouds first hit single "All your interwebs are belong to us" will be out this Friday Bitchez. Or maybe Monday. Soon though. Pretty sure.
Tough to have a cloud without a few rainy days.
Very strange things have been happening to everyone's Internet the past two weeks.
I read Indonesia's leading newspapers on-line because of one of my investments, and every time I do so, my browser turns pink, purple, and turquoise, while lions roar and fire alarms sound. Or a near equivalent.
Clearly, this is either the Iranians, Israelis, Bulgarians, BATS, operatives of the Koch Brothers, or a seventh grade honors class in Taipei working on a Science Fair project.
It's the joozz........
Can't wait until 30,000 drones come crashing out of the clouds.
A quick second shot on the cloudsource oracle aaaaannnd it's gone!
This is pointless. If you ran your AWS cloud in multiple zones, as you should, you would be online. This is some cheapskates that cut corners, and didn't run multiples. Same thing happens if you only have 1 server. Try Scalr, multiple zones, or something called PLAN B. It just shows these companies are run by asshat tech people who know nothing about redundancy, except when it comes to their business models.
Private clouds are fine.. but you cant outsource them... oh well... save a buck and let it all crash.. long scotch, candles and any real world activity
All of this cloud crap is being pushed by the technocrats and has a very sinister motive. If you just think about it from a criminal mind standpoint, then it is easy to see. We are dealing with a very sick criminal syndicate and the sooner we all realize this the sooner we can break free and root these sick people out of society.
I have rejected the 'cloud' services and have signed on with the pie in the sky servers.
Much better option.
Only if you live in Strawberry fields forever Disenchanted.
Cloud computing means nothing,this term is used by those who don't understand computer science.
http://www.youtube.com/watch?v=0FacYAI6DY0
I've audited the IT controls at one of the nation's largest cloud providers. In fact I just got done with the audit a couple of weeks ago.
This is an issue. But, it is overblown in this article. You are not going to take down the entire Internet because they are on the cloud. That is preposterous for several reasons.
Each large cloud provider has multiple datacenters spread across the world. While complete replication of each site is not likely, there are ways to failover to other sites especially if customers opt for this option in their contracts.
Generally it is stated in the contract that customers are responsible for their own DR/BCP (Disaster Recovery, Business Continuity Planning). The providers can provide that service upon a fee, as well. Almost all customers elect to take their chances, and in many cases, they don't make their own arrangements adequate to dealing with a site down issue. Whether they are on the cloud or in a regular datacenter (theirs or someone elses), the issue of site down is the same and the cost of replication at a cloud provider can often be cheaper than if they do it themselves. In other words, there is no difference in DR/BCP in a cloud provider than by running your own datacenter, and it can often be done cheaper through the cloud provider.
The reality is, most customers are ill-prepared for site down situations and it is their own fault, not that of the cloud providers. Sad, but true.
Most cloud providers use virtual machines (VMs) running on Linux/Unix to run their boxes, and both the server and the data can be striped across multiple boxes. This offers some fault tolerance, but again not complete replication. The good news is that with proper backups, you can bring back up a VM in relatively short amount of time, if the files are backed up. Most VMs are simply a set of files which can be backed up to a hotsite or to tape. If these companies did not pay for backup services at the provider, and do not provide their own, then it is their own damn fault and not that of the cloud provider.
For the cloud provider to be truly at fault, they would have to have multiple site down failures at the same time, which I have never heard happening. Natural disasters do not strike multiple cities all at once, unless God decided to flood the world and start over. And then none of it would matter anyway.
For power issues, many sites who are higher tier (such as government compliant) have multiple grid connections in the event one power provider goes down. They also should have generators and UPS on site to mitigate any loss of data or connectivity. A world class datacenter has many ISP connections coming into their network. The good ones (IE those that host government and large commercial customers) have over 100 ISPs coming into their datacenters. In addition, they build a dual tier infrastructure. If a DDOS (distributed denial of service), Internet connectivity, or power issue affects one side of the network, the fully real-time redundant infrastructure would simply instantly fail over to circuit two. In most cases, both circuits are live at the same time and the failover is seamless to the users. In most cases both circuits are live and load balanced at all times. Each side can handle all traffic if neccessary.
I could go on, but you get the idea. The major issues in cloud are not an Internet shutdown. They are more likely:
-Data remnance left on disks and used by new customers (easily dealt with by encryption and crypto-shredding)
-Encryption of data transfers between host and client (almost always the client's issue and not the providers)
-E-discovery, IE: do you have to respond to government subpeona on certain issues (in most cases, providers have ways of refusing government data requests where mutitple tenant data could be compromised)
-Security mechanisms to deal with hackers and advanced persistent threats (APTs) (this can be dealth with by a Security Governance function and good IT Security people)
-some other IT issues that typically are solved by proper planning
I have been in IT for 13 years in various roles, and the cloud is just a mechanism that has the same potential for failure as any other computer system. It is more complex to implement, but with that complexity comes more speed and flexibility. If the customers plan appropriately, the cloud can be cheaper, easier, and more fault tolerant than running their own datacenters.
mega-uber-cloud will crash, and the entire internet will be down
and to think the internet was created with immunity to things like this
but it is not the problem with using coulds, it is the problem with everybody using the ONE cloud