Johan De Gelas's picture

Datacenters have been reducing their Power usage effectiveness (PUE) for years now. It has become an important factor, which can generate a lot of good PR. The PUE is defined as being equal to the sum of facility power and IT equipment power divided by IT equipment power. Facility power got most of the attention, but while datacenters have been reducing their PUE from roughly 2-4 to 1.2-1.5, the customers of the datacenters hardly know where to start. Yes, despite all the talk about green IT, few companies with large outsourced IT infrastructures have been able to reduce the energy bill. Server sprawl has been slowed down but was not stopped by virtualization and few people have taken the time to study the problem. So we felt it was to time for the Sizing Servers Lab to step in and do our share. 

On the first October we started with our IWT approved project "Saving Energy in the IT infrastructure". We will try to lower the energy consumption inside the rack at 5 levels:

  • Software tuning in both the development stage as in the configuration phase.
  • OS power management optimisation (ACPI)
  • Optimal hardware choice
  • Optimal virtual cluster settings (DRS for example)
  • Optimal physical cluster settings

No less than 17 flemish companies and institutions decided to partner with us to battle this problem:

  • Combell, one of the largest hosting firms in the Benelux is a big customer in many datacenters. 
  • UZ Gent and Agfa however have their own large datacenters and have thus control over both the IT equipment as the facilities.
  • Racktivity is the innovator behind the Energyswitch products. The Energyswitches allows us to measure energy accurately and helps us to control energy consumption. 
  • A-server is the company of the SmartStyle Office, an all in one office solution. We work with A-server to further improve the energy efficiency and performance of their product.
  • D&N is our partner in "Green software". Well optimized software can reduce the energy consumption by 5x and more. Thanks to Intel, we now have even a software tool that allows to measure energy efficiency of software.  
  • Hexios and Savaco are capable integrators which allow to understand what companies actually expect from their infrastructure.
  • A.N.S. sells IT equipment material and is thus an another interesting source of market information.  
  • RIS is a local service and hosting provider that is constantly seeking to improve their efficiency.
  • Ecopuur is driven by ecological and energy saving values and wants to apply the same philosophy to its IT infrastructure.
  • EnergyICT is a world renown energy management company. 
  • The datacenter departement of Ingenium consist of the seasoned architects of several datacenters. Being one of the first European "Code of conduct for data centers" endorsers, Ingenium offers us a different way to look at the energy consumption problems of the datacenter.
  • If we save a lot of energy inside the rack, these savings must also result in facility savings. That is the task of datacenter owner LCL: helping us understand the impact that the IT equipment has on the rest of the datacenter. 

VITO's Energy technology departement has been researching green technologies for a long time and University Gent research group ELIS is specialized in the hardware/software interface.

An impressive and long partner list means very little without some realworld results. We are of course at a very early stage of the research, but you can already find some of the bits of our research here. By custom tuning the power management of ESXi, we could sometimes significantly lower the power consumption at low/idle load:

Graph Power management tuning

 

Notice how AMD's newest "Interlagos Opteron" server platform consumes 20% (!) less thanks to a customized power management policy. We will report more as our research starts to produce more and more results.

Johan De Gelas
13/12/2011 - 18:43
Philip Dubois's picture

On the 2nd of March, we at the Sizing Servers Lab organized the "SME Datacenter Automation" seminar at VOKA. One of the topics included was about openQRM and how to set it up. With this blog we would like to publish our slides and the hands-on document, so you yourself can try it out.

openQRM is not just yet another software package offering cloud computing, but aims to achieve the creation of a complete dynamic datacenter. It profiles itself as a "Open Source Datacenter Management and Cloud Computing Platform", designed to fully automate datacenters in a scalable manner. It combines physical and virtual machines in a single management console. openQRM integrates with all mainstream virtualization technologies (Xen, KVM, Citrix XenServer, VMware ESXi, ...) and supports transparent Physical-to-Virtual, Virtual-to-Physical and Virtual-to-Virtual migrations.

For more information of this product, please visit the openQRM Enterprise website.

Philip Dubois
10/03/2011 - 14:46
Johan De Gelas's picture

The long period without updates on the english part of our site might suggest that we are only focused on our Flemisch activities. Nothing could be further from the truth! At the end of last month we have given a seminar on "hybrid clouds and advanced virtualization" in Karlstad. This event was organised by our Swedish partner Compare, a joint organization of the ICT industry in Sweden. Compare is supported and supports about 100 companies in Sweden and has lot of datacenter expertise.

If you like to see the presentations we have given, you can find them here. The content was mostly an updated version of what we presented on our last themeday.  We are currently looking into a collaboration around "green servers" or how you can improve the power management of what sits insides your rack. Traditionally, datacenters have focused on everything around the rack: UPS, power conversion, Cooling efficiëncy and so on. We are actually looking on how you can save energy without lowering the performance of your servers. To be continued.

Below you see us visiting Compare's ultra modern experimental datacenter where the newest cooling and UPS technologies are tried out long before they go into production.

Our team was also present at the Devopsdays in Hamburg, Liz went to the Intel Innovation Conference in Dublin and last but not least we exchanged ideas at VMworld Europe (Copenhagen, Danmark). You can find our thoughts and impressions about hybrid clouds here.

Johan De Gelas
13/10/2010 - 14:32
Liz van Dijk's picture

After about a week of heavy experimenting, the new website is finally online!

"New website?", you say. "Seems like nothing's changed..."

Nothing is further from the truth however. Though our last website was created in Flash by a very talented Multimedia-intern, we chose to keep only the design, but recreate the entire system in a Drupal-based environment, making the site a lot easier to maintain.

At this point, we consider the website to be in beta, so it's possible some esthetic problems might pop up, though we'll try to work these out as soon as possible. Also, we are always open to suggestions, the new Commenting-function allows people to directly comment on our Blog-messages.

Liz van Dijk
14/10/2009 - 13:12
Liz van Dijk's picture

After a bit of a hiatus in properly maintaining this site due to various maddening deadlines and the stress that comes with finding financing for the coming few years (we succeeded, by the way!), starting today, this site will be receiving weekly updates, so stay tuned!

As for today's update, it's time to share a bit of info on the things we learned at this year's VMworld Europe in Cannes. As we 4-manned the event, we were able to attend quite a large amount of sessions. A full-sized article on that is coming up on Anandtech, but for our own website, we'll put up a couple of our notes to start out with:

Performance Best Practices (a session by Scott Drummonds)
We didn't learn quite as much as we'd hoped from this session, but its contents are nonetheless important considerations for anyone hoping to virtualize their data center. We'll list the most interesting pointers down here:

  • Any application that spends most of its time using the CPU and not much of anything else will barely feel the impact of virtualization. Performance is impacted primarily by applications that require a lot of time spent switching to kernelmode.
  • The highest measured network speed using ESX is 16Gb/s
  • ESX is able to handle at most about 100k IOPS and 600 disks in total.
  • Today's CPU's are much better equipped to handle virtualization than those of a few years ago. Consider upgrading for a very large performance boost.
  • The use of jumboframes is recommended for every NIC. Separate NICs should always be used for interserver communications within the data center, like VMotion.
  • Disk caching and properly distributed disks are an important factor for storage performance. The newest version of ESX will improve a lot on iSCSI performance.
  • VMware recommends always using VMFS with 64k boundaries. This is the standard setting in vCenter. If you are not using vCenter, remember to set it yourself.
  • Use an OS that causes less interrupts, and disable every device you don't need.
  • Enable large pages for the TLB, to reduce cache misses.
  • Applications that do not scale well on multicore platforms natively might do well in a multiple VM-deployment. This way, each app can perform at its most optimal setup.

If you're interested in learning more, stay tuned! We'll be updating this blog regularly with interesting tidbits of information.

Liz van Dijk
27/03/2009 - 16:44
Syndicate content