Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>I would go for decentralized solution. </p> <p>Every node <em>picks</em> (not given) same amount of work from center. After some run every node is able to deside for <code>itself</code> an average power of calculation and communicate it with others. </p> <p>After all every node will have a table of every node's average calc power. Having this information (could be even persistant,why not?) each node can deside to "ask" some other node with more power to delegate a stuff to it by signing a contract.</p> <p>Before every process start every node have to make broadcast signal about: "I start doing X". One time finished always broadcast: "I finished X". </p> <p><strong>Well, it's no so easy</strong>, cause there will be case when you begin job, after your hard disk failed and you will never finish it. Others, especially those ones who are waiting a result from you should figure out this and pick from the basket your job and begin the stuff from the beginning. Here come "ping" technique with timer. </p> <p>Bad: The first tuning time can take non indifferent amount of time. </p> <p>Good: You will have <em>almost</em> fault tolerant solution. Leave them for a week, and even if some of nodes fail your grid still alive and does its work. </p> <p>Many years ago I did something like this and with pretty good results. But it wasn't definitely on such large scale as described by you. And scale, actually, makes a difference.</p> <p>So the choice is up to you.</p> <p>Hope this helps.</p>
    singulars
    1. This table or related slice is empty.
    1. POLoad-balancing in parallel processing application
      singulars
      1. PTQuestion
    1. PTAnswer
    1. This table or related slice is empty.
    1. USTigran
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    1. COThe scalability and robustness of this solution is *terrible*. Broadcasts scale horribly. And delegation may work well to balance differences in processing power, but does nothing for the case where a node drops offline.
      singulars
      1. PO
      1. USBen Voigt
    2. CO@Ben Voight: One of the problem is a startup tuning time (as mantioned in aswer), while nodes are going to balance in the grid. What about fault tollerance: In my solution nodes were going to lokup in central place to get a job to do. Checking of if the job completed or not is kind of tricky too. By the way, as I mantioned my solution was on definitely smaller scale then that one asked in question. If you have an experience with such kind systems, I personally, would curiose to read your answer/comments on subject.
      singulars
      1. PO
      1. USTigran
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload