1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Cloud computing gone bad? Amazon servers down...

Discussion in 'General Hardware' started by johnspack, Apr 21, 2011.

  1. johnspack

    johnspack

    Joined:
    Oct 6, 2007
    Messages:
    4,382 (1.68/day)
    Thanks Received:
    865
    Location:
    Nelson B.C. Canada
  2. Kreij

    Kreij Senior Monkey Moderator Staff Member

    Joined:
    Feb 6, 2007
    Messages:
    13,881 (4.88/day)
    Thanks Received:
    5,616
    Location:
    Cheeseland (Wisconsin, USA)
    Personally, I think the "cloud" is a disaster just waiting to happen.
     
    95Viper says thanks.
  3. hellrazor

    hellrazor

    Joined:
    Feb 18, 2010
    Messages:
    1,580 (0.91/day)
    Thanks Received:
    319
    Everybody! Put all your eggs in one basket that likely never gets updated!
     
  4. FordGT90Concept

    FordGT90Concept "I go fast!1!11!1!"

    Joined:
    Oct 13, 2008
    Messages:
    13,950 (6.25/day)
    Thanks Received:
    3,786
    Location:
    IA, USA
    Same. The only reason why businesses like Google, Amazon, and Microsoft are pushing it is because it gives them information about you and information is power. I like to keep all my data local, secure, and with multiple backups, thank you very much.
     
    95Viper says thanks.
    Crunching for Team TPU
  5. streetfighter 2

    streetfighter 2 New Member

    Joined:
    Jul 26, 2010
    Messages:
    1,658 (1.05/day)
    Thanks Received:
    732
    Location:
    Philly
    What's "ROF sim"?
    It's always sounded that way to me . . . On the other hand I was under the impression that the "cloud" is supposed to negate SPOF with various failsafe mechanisms (such as failover).

    I'm disappointed that EC2 went down like dominoes.
     
  6. Kreij

    Kreij Senior Monkey Moderator Staff Member

    Joined:
    Feb 6, 2007
    Messages:
    13,881 (4.88/day)
    Thanks Received:
    5,616
    Location:
    Cheeseland (Wisconsin, USA)
    I guess my question in this case would be, "Where is the "cloud" redundancy?"
    If Amazon's Virginia site went down, why wasn't all traffic immediately rerouted to another cloud location where the data was replicated in real time?

    It's not being replicated? Nice, that's network management 101 stuff. :shadedshu

    If I was a cloud customer who is paying for terabytes of data to be available to my customers and it became unavailable, I would be .. well .. perturbed.

    @SF2 : ROF is Rise Of Flight.
     
    Last edited: Apr 22, 2011
    streetfighter 2 says thanks.
  7. johnspack

    johnspack

    Joined:
    Oct 6, 2007
    Messages:
    4,382 (1.68/day)
    Thanks Received:
    865
    Location:
    Nelson B.C. Canada
    Man, it's still down. I just bought a plane, and someone gifted me one, which I haven't even seen yet. We are trying to finance the next major patch to our sim engine. This hacker crap has just destroyed $1000s of effort on our part. What crap.
     
  8. 95Viper

    95Viper

    Joined:
    Oct 12, 2008
    Messages:
    4,441 (1.99/day)
    Thanks Received:
    1,623
    Location:
    στο άλφα έως ωμέγα
  9. Goodman

    Goodman

    Joined:
    Jun 13, 2009
    Messages:
    1,519 (0.76/day)
    Thanks Received:
    324
    Location:
    Canada/Québec/Montreal
    Was it ever anything different?
     
  10. bostonbuddy New Member

    Joined:
    Apr 14, 2011
    Messages:
    381 (0.29/day)
    Thanks Received:
    39
    the cloud always seemed like a scam so the companies could make more money.
     
    W1zzard says thanks.
  11. Funtoss

    Funtoss New Member

    Joined:
    Mar 4, 2011
    Messages:
    588 (0.43/day)
    Thanks Received:
    60
    Location:
    New Zealand
    what is cloud computing? o_O
     
  12. micropage7

    micropage7

    Joined:
    Mar 26, 2010
    Messages:
    6,081 (3.57/day)
    Thanks Received:
    1,413
    Location:
    Jakarta, Indonesia
    please read this

    cloud is pretty nice concept that said you can access your data no matter where you are and what device you are on
    but for today i guess theres alot that need to do from traffic, apps, security and so
    maybe in the future we could, but for today i guess its still need a long way to go
     
  13. Easy Rhino

    Easy Rhino Linux Advocate

    Joined:
    Nov 13, 2006
    Messages:
    13,499 (4.60/day)
    Thanks Received:
    3,330
    It was only a matter of time before we saw this. I wonder how this will effect IBM and Intel's sales.
     
    Crunching for Team TPU
  14. W1zzard

    W1zzard Administrator Staff Member

    Joined:
    May 14, 2004
    Messages:
    15,007 (3.90/day)
    Thanks Received:
    11,877
    if people stop using the cloud [for hosting], many of them will use dedicated servers, which means installed server count goes up
     
    95Viper says thanks.
  15. johnspack

    johnspack

    Joined:
    Oct 6, 2007
    Messages:
    4,382 (1.68/day)
    Thanks Received:
    865
    Location:
    Nelson B.C. Canada
    Well, the ROF master game server is back up thank god, and so is the store. 777 Studios was having a big sale as Easter weekend is their biggest selling time of the year. They lost a load of money on this. They count on the revenue to continue building up the game engine, and we are waiting for some really cool new features expected in the next patch. This really screwed a lot of people....
    Edit: our main forum is still down however.
     
  16. Kreij

    Kreij Senior Monkey Moderator Staff Member

    Joined:
    Feb 6, 2007
    Messages:
    13,881 (4.88/day)
    Thanks Received:
    5,616
    Location:
    Cheeseland (Wisconsin, USA)
    True .. unless the hosting companies are renting space in the cloud for the data so they do not have to manage their own infrastructure.
    Best to ask if your hosting company is using their own servers for the hosting services.
     
  17. CJCerny

    CJCerny

    Joined:
    Nov 25, 2008
    Messages:
    855 (0.39/day)
    Thanks Received:
    249
    Location:
    Akron, OH
    I understand the frustration, but I think the frustration or fear of cloud computing is totally unmerited. You're an idiot if you think any system ever devised by man is infallible. You have to allow for downtime, both planned and unplanned, whether it be for cloud computing or for that shiny new Camaro you just bought--there is going to come a time when it just won't work when you need it to.

    At the company where I work, we rely on LogMeIn to remotely control and diagnose the machines in our customer base. This past Sunday, they made some kind of backend change to their product that completely killed off our ability to control about 1800 machines remotely. It's not evil or the fault of the "cloud". It just happens.
     
    Last edited: Apr 22, 2011
  18. Frick

    Frick Fishfaced Nincompoop

    Joined:
    Feb 27, 2006
    Messages:
    10,885 (3.41/day)
    Thanks Received:
    2,395
    Meh, this stuff can happen without the help of cloud computing anyway, so I don't think it's the fault of the Cloud. I like the cload as a concept.
     
  19. Easy Rhino

    Easy Rhino Linux Advocate

    Joined:
    Nov 13, 2006
    Messages:
    13,499 (4.60/day)
    Thanks Received:
    3,330
    The cloud concept is great for saving power and money, but as we just saw, if the system goes down IT ALL GOES DOWN. so where amazon was offering money savings through some redundancy they may have to add even more now negating any big saving.
     
    Crunching for Team TPU
  20. Kreij

    Kreij Senior Monkey Moderator Staff Member

    Joined:
    Feb 6, 2007
    Messages:
    13,881 (4.88/day)
    Thanks Received:
    5,616
    Location:
    Cheeseland (Wisconsin, USA)
    You're right. Unplanned outages can occur.
    I'll remember to tell my customers that due to a unforseen consequence and the fact that I failed to set up an adequate contigency plan to even give them partial service in the event of a catastrophic failure, that, "Hey ... shit happens."

    I'm sure they'll be understanding and quite forgiving.
     
    Easy Rhino says thanks.
  21. johnspack

    johnspack

    Joined:
    Oct 6, 2007
    Messages:
    4,382 (1.68/day)
    Thanks Received:
    865
    Location:
    Nelson B.C. Canada
    Heh, yeah I use logmein as well to remotely control a server on the us east coast. It's a paid job, not good pay, but it's work. If logmein went down, I've have a very serious problem....
     
  22. FordGT90Concept

    FordGT90Concept "I go fast!1!11!1!"

    Joined:
    Oct 13, 2008
    Messages:
    13,950 (6.25/day)
    Thanks Received:
    3,786
    Location:
    IA, USA
    From the mouth of the lion: Amazon S3 Availability Event: July 20, 2008

    And the single point of failure that brought the whole thing down was corrupt bits in non-errorchecking environment (it is now). The servers were giving incorrect information to each other so the other servers were trying to compensate for it. Eventually, the whole network of systems was bogged down by bogus overhead.

    It was a programming error.


    It doesn't all go down 99.7% of the time. That only happens when there is a failure in inter-server communication which is quite rare.
     
    Last edited: Apr 23, 2011
    Crunching for Team TPU
  23. Easy Rhino

    Easy Rhino Linux Advocate

    Joined:
    Nov 13, 2006
    Messages:
    13,499 (4.60/day)
    Thanks Received:
    3,330
    I understand how servers work. This crash was completely unacceptable. If you build up an entire server cloud system around "reliability, efficiency and flexibility" then it should never ever go down like that. There should have been checks in place or some kind of policy that requires the admins to check for irregular activity. That could not have possibly gone unnoticed for long but who knows...
     
    Crunching for Team TPU
  24. FordGT90Concept

    FordGT90Concept "I go fast!1!11!1!"

    Joined:
    Oct 13, 2008
    Messages:
    13,950 (6.25/day)
    Thanks Received:
    3,786
    Location:
    IA, USA
    It was like a virus. Once the problem took hold, it spread. The only way to stop it was to reboot the entire cloud. It shares a lot in common with that Northeast Corridor power failure in 2000 (I think). One transformer/database went down and the rest responded as they were programmed to but, instead of isolating the issue in both cases, they magnified it.

    You can't blame them for being human.
     
    Crunching for Team TPU
  25. Easy Rhino

    Easy Rhino Linux Advocate

    Joined:
    Nov 13, 2006
    Messages:
    13,499 (4.60/day)
    Thanks Received:
    3,330
    oh i dont blame anyone. and i like cloud technology. but if amazon thinks that their cloud business will one day bring in more revenue for them than sales on amazon.com (which is what their ceo says they want) then they better start adding redundancy and people better get used to paying a little bit more for it.
     
    Crunching for Team TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page