Slashdot Mirror

← Back to Stories (view on slashdot.org)

Google Launches Cloud Dataproc, a Managed Spark and Hadoop Big Data Service

Posted by samzenpus on from the new-player dept.

An anonymous reader writes: Google has a new cloud service for running Hadoop and Spark called Cloud Dataproc, which is being launched in beta today. The platform supports real-time streaming, batch processing, querying, and machine learning. Techcrunch reports: "Greg DeMichillie, director of product management for Google Cloud Platform, told me Dataproc users will be able to spin up a Hadoop cluster in under 90 seconds — significantly faster than other services — and Google will only charge 1 cent per virtual CPU/hour in the cluster. That's on top of the usual cost of running virtual machines and data storage, but as DeMichillie noted, you can add Google's cheaper preemptible instances to your cluster to save a bit on compute costs. Billing is per-minute, with a 10-minute minimum."

6 of 18 comments (clear)

  1. How to max out someone's billing by msobkow · · Score: 1

    So in order to max out someone's billing, just run a query that will take half a second or few once every ten minutes to make sure that "ten minute minimum" is applied throughout every hour of the day. :(

    --
    I do not fail; I succeed at finding out what does not work.
    1. Re: How to max out someone's billing by Anonymous Coward · · Score: 1

      And we alll know that self-hosted services never go down.

    2. Re:How to max out someone's billing by msobkow · · Score: 1

      The "cloud" still has the same issues people have complained about for years:

      - Usually no effective way to back up large data sets to local media
      - Usually no effective way to restore large data sets from local media
      - Unpredictable costs
      - Single point of failure: one service provider
      - "All or nothing" failures
      - Virtually impossible to switch providers
      - Performance limited by network bandwidth
      - Difficulty loading large data sets for initial deployment
      - At the mercy of the provider; often no way to prioritize deployments if repair/recovery are necessary
      - Security at the discretion of the provider
      - Nothing more than good old fashioned clusters, just hosted off-site with fancy new buzzwords

      --
      I do not fail; I succeed at finding out what does not work.
    3. Re: How to max out someone's billing by msobkow · · Score: 1

      <rolleyes>Of course you could try using the &lt;/&gt; explicitly...</rolleyes>

      --
      I do not fail; I succeed at finding out what does not work.
    4. Re:How to max out someone's billing by jonbally7631 · · Score: 1

      Nowadays Billing is tough and old way to pay on hand. You should have to pay online via card or something else. Pay for securing the data in the drive.

  2. dataproc by robi5 · · Score: 1

    How is it pronounced?