back to article HBase: Shops swap MySQL for open source Google mimic

Facebook isn't the only one swapping MySQL for HBase, the open source distributed database platform based on Google's BigTable. The Hadoopian HBase is now in play at several of the web's most recognizable names – including Adobe, Yahoo!, Mozilla, and StumbleUpon – as well as smaller operations looking to climb their way to such …


This topic is closed for new posts.
  1. Magnus_Pym

    It's funny y'know...

    ...getting older. You can reminisce about all the old Microsoft debacle. So, 13 years since they f***ed up with the Hotmail change over eh! who'd have thought it. 'course I can remember back in the day when win95 started supporting USB. oh the fun we had...

    1. Dave Murray

      Back in the day?

      Win95 is hardly back in the day. I remember having to put the DOS floppy back in the drive because the app I was using needed to access Now that was back in the day.

      1. This post has been deleted by its author

        1. Bob Foster

          "Incorrect DOS version"

          "Incorrect DOS version" , you were lucky!, I remember having to get up at three in the morning walk ten miles into work and put cards in't punch card reader!

          1. M Gale


            Why when I were a lad, we had to crawl around the whole bloody computer all day, yanking dud valves out. And when we got home, to our matchbox-size house with eighteen kids in it, our father would belt us black and blue, telling us we were lucky to be working with them newfangled valve things instead of relay switches.


            1. Chemist

              "..yanking dud valves out.."

              Oddly enough I was taught Physics by someone who had worked as a graduate student on the Manchester Baby. He had a photo of himself, stripped to the waist, working in a room stacked with chassis. No A/C of course and precious little ventilation

              1. paulc


                and he was finding real bugs as well...

            2. Allan George Dyer

              Valves? Valves!

              When I were a lad, we didn't have these fancy valves. When the system went down, we had to go to Wales a quarry out another set of bluestones. Then transport them back, balanced on dugout canoes. And kids these days just don't believe you!

              Mine's the one with the bronze helmet.

  2. Rinsey

    Ada Lovelace

    Now she was back in the day

  3. Anonymous Coward

    Pfft, changing the floppy?

    I remember having to hold the headphone jack on the speccy 48k at a specific angle whilst loading software, or it would fail to load. No matter how many times I replaced that cable, it always used to $^&&*( fray :/

    1. M Gale

      This is why I liked the BBC Micro.

      Loaded data from the cassette in discrete blocks. If you fucked up one block, just rewind the tape and replay it. No need to start the whole *&$^ing 15-minute loading sequence all over again.

      Aye, thems were the days. Young 'uns these days, don't know they've been born!


    2. paulc

      milk carton on heatsink

      had to put a milk carton on the heatsink and a thick rubber band holding the RAM pack tight against the expansion slot so I could get a program typed in and saved off on my ZX81...

      ah the joys of having the puter crash or restart halfway through cos the voltage regulator had overheated or else the RAM pack had wobbled...

      ended up building a custom case for it in the end with the heatsink bolted to the aluminim case and the RAM pack screwed hard into place...

      and the good old days when printers came with a cable and a massive manual complete with enough info to hand code your own printer driver if needed

  4. John Latham

    Cut and paste from a webpage?

    Surely that would be a rather serious security flaw.

  5. Destroy All Monsters Silver badge

    So the bottom line is....

    HBase is useful in some cases.

    1. Ian Michael Gumby

      Some cases?

      HBase is useful in a lot of cases.

      Lets talk scalability.

      HBase can scale in linear or near linear terms. You can hash your key and get a fairly random and equal distribution across your region servers. This is critical when you consider that if you have a 10 node cluster or a 100 node cluster, pulling data (a simple get() ) will occur in a consistent time. That is to say fetching a customer record where cust_id = XXX will take the same amount of time regardless of the amount of data in the cluster.

      Looking at relational databases, Oracle RAC? Anyone trying to scale that past 8-10 machines?

      DB2? IBM is still improving it with stuff they inherited from Informix.

      IBM was foolish to kill or rather starve Informix's XPS which also boasted near linear scalability back in the 90's.

      But here's the kicker. HBase is 'free'. Although you would want to get a support contract if you were using it in a commercial production environment. So the cost of building out 100s of TB or a PB cluster is primarily the cost of the hardware.

      You can't get that scalability and performance from mySQL or anything else.

      Considering that HBase is < 5 years old, the technology is improving and in the last year, stability has increased.

      So yeah, I'd say HBase has its place.

      You just have to understand what it is and how to use it.

      1. Anonymous Coward
        Anonymous Coward

        Not lots, but some...

        ..."some" being scenarios where scalability is an absolute requirement. Anything that requires *real* ACID compliance better stick with MySQL. It's the MongoDB v. RDBMS argument all over again (HBase will guarantee a flush to the OS cache, but not a write to disk). As the HBase documentation points out, if you loose a data centre you loose durability. To be honest I have no idea whether HBase supports any DR scenarios or not. Even so, if my other choice was Oracle I'd become an HBase bigot.

        1. Ian Michael Gumby


          Since HBase 0.89 there is code for replicating between clouds. I haven't played with it... yet... and today (yesterday by the time you read this...) HBase version 0.90 was released.

          So you can use log shipping to get across the loss of a data center.

          For most people using mySQL, they do not have more than one instance in multiple data centers so that is a moot point.

          Trying to do DR or back up and recovery is a major issue when you're dealing with 100+TB because you can't easily take backups ;-)

          But to your point about HBase not being ACID compliant, you're right. Its not the right tool for that job. And yes, you can use Hadoop/HBase to feed in to your RDBMs system.

          1. Anonymous Coward

            @Ian Michael Gumby

            > Since HBase 0.89 there is code for replicating between clouds.

            What in the fsck does "replicating between clouds" mean? Does it copy water molecules in suspension from one cumulus to another?

            Oh my! You are full of cloud-marketing speak.

        2. Lennart Sorensen

          Well no not mysql either.

          If you want a proper database you would of course be using postgresql, not mysql. I have never been able to understand this widespread fascination with mysql.

  6. Anonymous Coward
    Thumb Up

    Strapline says it all ...

    "Microsoft doesn't want it. But everyone else does"

    Ideal. Anything MS doesn't want is definitely worth looking at.

  7. Sam Penny

    First microsoft use of open source?


    The HBase-based Powerset was folded into Bing, making the search engine one of the first "shipping" Microsoft product to actually include open source code.


    What about the BSD Networking stack, wasn't that the basis of all windows' networking stacks from way back in the days of windows 95 (or maybe even 3.11)?

    Cheers & God bless

    Sam "SammyTheSnake" Penny

  8. D. M
    IT Angle

    back to old days

    Since when win95 is considered old days? I'm relatively young, I remember the days you use tape to load software, Hugh 5' single side floppy, the time before MS DOS and IBM PC.

    1. david bates


      1) 5 1/4" floppies (no idea where you get 5 FOOT floppies from...).

      2) They were no more than medium size - I remember (just) 8" floppies...

  9. David Kelly 2

    "swapping MySQL for HBase"?

    Is it my American English which defines the evaluation precedence of the above phrase to be saying MySQL is replacing HBase rather than the opposite stated in the article?

    The "for HBase" suggests HBase is there first.

    I would have better understood "swapping MySQL with HBase", but would have expected an article on making the two interchangeable.

    "Replacing MySQL with HBase" would better say what was intended.

    1. Anonymous Coward
      Anonymous Coward

      wrong its your american english.

      If we said people are swapping this for that. One would understand this to mean that this is in place and that that is the replacement for this.


  10. Anonymous Coward
    Anonymous Coward

    Could we clarify?

    I am not convinced that the author of this article understands the difference between a distributed non-relational database and a relational database with limited support for distributed scenarios.

    The right tool for the job, horses for courses, and all that. If the article was trying to intimate that one product is better than the other in an absolute sense, or that they are directly comparable, the wrong end of the stick was being held.

    On the other hand if I have underestimated the author's knowledge of the technology being discussed, I sincerely apologise, but I still think that this article could have done a better job of reflecting that knowledge.

This topic is closed for new posts.

Other stories you might like