back to article So you want to build the next Google. Who ya gonna call? Er, Big Blue?

IBM has announced a new version of its Platform Resource Scheduler (PRS), which lines up jobs and resources in mammoth OpenStack Havana environments. In doing so, Big Blue hopes to give enterprises a shot at achieving the same levels of efficiency as Google's highly tuned servers. Though the tech competes against VMware's …

COMMENTS

This topic is closed for new posts.
  1. Muppetry

    Tibco (via their data synapse acquisition) were able to do something similar a long time back for either virtualised or bare metal resources. In essence it was an extension of their meta-scheduler for HPC applications that got extended to cover transactional applications. It didn't plug into a cloud orchestrator, but straight into the hypervisor, but there was nothing stopping it being tied to any kind of orchestration / automation with a bit of time & effort.

  2. Richard Taylor 2

    Meaningless

    "Google's Borg system is rumored to have been so good at this task juggling act that it saved the ad-slinger from building an entire data center."

    So a rumour of an advantage that you are incapable of quantifying the benefit (was this a one machine data centre, was it one of 3 of 5 or of 1000?) Classy fact checking

  3. Uncle Ron

    Whether or not...

    I too comment on the "Google's Borg system is rumored to have been so good at this task juggling act that it saved the ad-slinger from building an entire data center" statement. Even if the implication is true, that Google saved a data center, I say, So What? A data center is only 10's of millions of dollars and a few handfuls of employees. A literal drop in the Google bucket. The power and cooling for all those coders probably exceeds that of a data center...

    I still feel that throwing hardware at a problem, when today's hardware is -very- cheap, makes more sense than developing more and more complex software that always leads to lots of unintended consequences. As a matter of fact, if 90% of my jobs (which is probably true) require an ARM processor with 4GB of memory, why not have a million ARMs with channel access to lots of segmented memory?

    As Albert Einstein said, "Everything should be made as simple as possible, but no simpler.”

  4. Gregg McKnight

    Your perspective is typical for most software developers, especially ones with enterprise development experience. You see the cost of your laptop or desktop as cheap... But the cloud changes this perspective significantly. A Google datacenter costs about $10M per MegaWatt (MW). Most Google DCs are at least 50MW. That puts the datacenter costs at about $500M, not 10s of Millions. Furthermore this does not include servers, network, and storage. These systems will cost Google about another $10M/MW. A typical Google DC will run well over $1B. Not cheap by anyone's standards... In the cloud, software is constantly redeveloped to make more efficient use of the HW, not the other way around...

This topic is closed for new posts.

Other stories you might like