ad: CQMM-1

QRZ Temporarily Shut Down by Amazon Web Services Outage

Discussion in 'Amateur Radio News' started by AA7BQ, Oct 22, 2012.

Thread Status:
Not open for further replies.
Page 2 of 4
ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: Left-3
ad: abrind-2
ad: Left-2
ad: Radclub22-2
  1. W8QZ

    W8QZ Ham Member QRZ Page

    So, this is the 'in the cloud' future of computers? Hmmm . . . . . .
     
    W8QZ, Oct 23, 2012
    #16
  2. KI4SP

    KI4SP Ham Member QRZ Page

    very strange, so was the Army AKO mail (Army Knowledge Online)...........
     
    KI4SP, Oct 23, 2012
    #17
  3. AF6LJ

    AF6LJ Ham Member QRZ Page

    colocation of equipment would be the best way to, it's also the most expensive. Even still there are issues, like you are not always to access your gear to preform repairs or upgrades, that brings up the other problem. Your gear and the cost of upgrading it maintaining it, so the best option for some businesses is the cloud. All the cloud comuting providers have outages, Amazon just happens to be one of the largest...


    It is for many but not all. These days in online gaming there are a few big companies who host servers and provide support however for many of the games the actual game server is a dedicated peace of hardware sometimes owned by a couple of people colocated at a small to medium sized ISP. There are many companies who still have colocated hardware. Healthcare is moving away from this model and is in the cloud these days.... My roommate's office has had to deal with outages as everything but their appointment book is in the cloud, doctors notes everything....
    I don't know of anybody who really can stand up and say this is a good thing, expect for the fact everything is backed up off site. That is one good thing about the cloud...
     
    AF6LJ, Oct 23, 2012
    #18
  4. KY5U

    KY5U Ham Member QRZ Page

    The computer industry needs a red herring buzzword to chase every so often to get the sleazy sales types excited. "The Cloud" is the latest. For some uses cloud storage may make a difference, but with reliability and security issues it has a long way to go to be a replacement for local storage of data.
     
    KY5U, Oct 23, 2012
    #19
  5. N0AMT

    N0AMT Ham Member QRZ Page

    Boy, this has OPSEC failure written all over it. Please consider what you're posting in the open for all to see here. If what you say is true, do you really think everybody who comes to QRZ has a need to know? Should we really be telling the world?

    Not possible. Every provider goes down from time to time, often from problems outside of their control. Cloud or no cloud.

    Cloud or no cloud, I'd urge you to find me ONE provider, who can provide all that AWS does for QRZ with absolutely zero downtime guaranteed. Nothing exists.


    Folks keep in mind, the internet is built on top of hardware. From time to time (as evident in our own hobby) hardware fails. This could happen to your small town ISP and bump a few people offline, or it could happen to Amazon, or even Amazon's backbone provider, and bump more than a few websites offline.

    Some of the other websites affected:
    Reddit, Netflix, Pinterest, Instagram, Github

    If AWS is good enough for them, it should definitely support QRZ.
     
    Last edited: Oct 23, 2012
    N0AMT, Oct 23, 2012
    #20
  6. AA0V

    AA0V Premium Subscriber QRZ Page

    Capital One is still down. I called Sunday, they said "It was on their end" Makes me wonder??
     
    AA0V, Oct 23, 2012
    #21
  7. K5CO

    K5CO Ham Member QRZ Page

    And no one thinks it comes from Persia?
     
    K5CO, Oct 23, 2012
    #22
  8. AA7BQ

    AA7BQ QRZ Founder QRZ HQ Staff QRZ Page

    Having been Manager of World Wide Datacenter Operations at a multi-billion dollar company, I do have some background in servers, hardware, and their administration.

    QRZ had its own hardware and ran in colocation facilities for about 15 years. The hardware we had was very reliable and the only component that ever went down was a router. It was pretty stable, but, as another poster mentioned, not necessarily cheap to operate. It also meant that if I was on vacation, thousands of miles from home, that the only recourse in the event of a hardware failure was to fly home and fix it. There's no telling how long such an outage would have lasted and, hopefully, spare hardware would be available to fix the problem. You can't drive down to the computer store and pick up a SAS drive, or a motherboard for a Sun server.

    In the cloud, I don't know what the hardware looks like, nor do I care. The entire QRZ service is compressed into a set of images that can be un-zipped and started up on a fresh server instance in a matter of minutes. I've spent weeks fine-tuning the images to make that possible. This is true hardware independence and it's a wonderful thing.

    The cloud at Amazon is very large and complex. Just consider MTBF's for example. Let's say that a new disk drive has a 200,000 hour MTBF. In a room with 2000 disk drives, you can expect a hard disk failure about every 100 hours. In addition, these services employ SAN, or Storage Area Networks that are fiber based and go through fiber routers.

    The bottom line is that redundancy is available, at a price. QRZ could be spread out over several Amazon datacenters around the globe and configured for automatic fail-over as well as database replication. This could easily triple our data processing costs which already run in the thousands of dollars per month. The increase in price would specifically address the inconvenience of downtime that in most cases is less than 12 hours. Yesterday's outage lasted perhaps 4 or 5 hours. Do you seriously think it's worth it to spend several thousand dollars per month to avoid such a small inconvenience? This isn't a banking or safety-of-life operation. There is never anything on QRZ that can't wait a few hours to resolve.

    Just consider that 24 hours of aggregated outage per year (which we haven't yet seen) still puts us in the 99.99% availability bracket (1/365 = 0.0027). With numbers like those, I consider the operation to be a smashing success!
     
    AA7BQ, Oct 23, 2012
    #23
  9. AF6LJ

    AF6LJ Ham Member QRZ Page

    Ah Yes the Ball And Chain aspect to managing your own hardware.
     
    AF6LJ, Oct 23, 2012
    #24
  10. K9ASE

    K9ASE XML Subscriber QRZ Page

    I wasn't sure if it was me or the Zed anyway.:) I played on 10 meters and fixed a neigbors car that broke down.
     
    K9ASE, Oct 23, 2012
    #25
  11. KF7PCL

    KF7PCL Ham Member QRZ Page

    I strongly suggest you don't get so twisted up about statements like this.
    A. It was unproven
    B. Even if it was true and he said it, it is not your problem
     
    KF7PCL, Oct 23, 2012
    #26
  12. N0AMT

    N0AMT Ham Member QRZ Page

    I fundamentally disagree with this statement entirely. It most certainly IS my problem, and it should concern everyone when people (correct or incorrect) take it upon themselves to share security information.
     
    Last edited: Oct 23, 2012
    N0AMT, Oct 23, 2012
    #27
  13. KF7PCL

    KF7PCL Ham Member QRZ Page

    The people that constantly babble about "national security" tend to make me simply face palm.
     
    KF7PCL, Oct 23, 2012
    #28
  14. WA6MHZ

    WA6MHZ Ham Member QRZ Page

    WOW! Looks like we are going to have a QRZed FIST FIGHT!!!
     
    WA6MHZ, Oct 24, 2012
    #29
  15. KA9JLM

    KA9JLM Ham Member QRZ Page


    I hope it is on Youtube.


    Really QRZ downtime is not all that bad.
     
    KA9JLM, Oct 24, 2012
    #30
Page 2 of 4
Thread Status:
Not open for further replies.

Share This Page

ad: cq2k-1