Tuesday, April 30. 2013A Smarter Algorithm Could Cut Energy Use in Data Centers by 35 Percent
-----
By David Talbot on April 16, 2013
Storing video and other files more intelligently reduces the demand on servers in a data center.
Worldwide, data centers consume huge and growing amounts of electricity. New research suggests that data centers could significantly cut their electricity usage simply by storing fewer copies of files, especially videos. For now the work is theoretical, but over the next year, researchers at Alcatel-Lucent’s Bell Labs and MIT plan to test the idea, with an eye to eventually commercializing the technology. It could be implemented as software within existing facilities. “This approach is a very promising way to improve the efficiency of data centers,” says Emina Soljanin, a researcher at Bell Labs who participated in the work. “It is not a panacea, but it is significant, and there is no particular reason that it couldn’t be commercialized fairly quickly.” With the new technology, any individual data center could be expected to save 35 percent in capacity and electricity costs—about $2.8 million a year or $18 million over the lifetime of the center, says Muriel Médard, a professor at MIT’s Research Laboratory of Electronics, who led the work and recently conducted the cost analysis.
So-called storage area networks within data center servers rely on a tremendous amount of redundancy to make sure that downloading videos and other content is a smooth, unbroken experience for consumers. Portions of a given video are stored on different disk drives in a data center, with each sequential piece cued up and buffered on your computer shortly before it’s needed. In addition, copies of each portion are stored on different drives, to provide a backup in case any single drive is jammed up. A single data center often serves millions of video requests at the same time. The new technology, called network coding, cuts way back on the redundancy without sacrificing the smooth experience. Algorithms transform the data that makes up a video into a series of mathematical functions that can, if needed, be solved not just for that piece of the video, but also for different parts. This provides a form of backup that doesn’t rely on keeping complete copies of the data. Software at the data center could simply encode the data as it is stored and decode it as consumers request it. Médard’s group previously proposed a similar technique for boosting wireless bandwidth (see “A Bandwidth Breakthrough”). That technology deals with a different problem: wireless networks waste a lot of bandwidth on back-and-forth traffic to recover dropped portions of a signal, called packets. If mathematical functions describing those packets are sent in place of the packets themselves, it becomes unnecessary to re-send a dropped packet; a mobile device can solve for the missing packet with minimal processing. That technology, which improves capacity up to tenfold, is currently being licensed to wireless carriers, she says. Between the electricity needed to power computers and the air conditioning required to cool them, data centers worldwide consume so much energy that by 2020 they will cause more greenhouse-gas emissions than global air travel, according to the consulting firm McKinsey. Smarter software to manage them has already proved to be a huge boon (see “A New Net”). Many companies are building data centers that use renewable energy and smarter energy management systems (see “The Little Secrets Behind Apple’s Green Data Centers”). And there are a number of ways to make chips and software operate more efficiently (see “Rethinking Energy Use in Data Centers”). But network coding could make a big contribution by cutting down on the extra disk drives—each needing energy and cooling—that cloud storage providers now rely on to ensure reliability. This is not the first time that network coding has been proposed for data centers. But past work was geared toward recovering lost data. In this case, Médard says, “we have considered the use of coding to improve performance under normal operating conditions, with enhanced reliability a natural by-product.”
Personal comment:
Still a link in the context of our workshop at the Tsinghua University and related to data storage at large.
Posted by Patrick Keller
in Architecture, Science & technology, Sustainability
at
07:40
Defined tags for this entry: architecture, code, data, energy, information, science & technology, sustainability
Monday, April 22. 2013A new way to report data center's Power and Water Usage Effectiveness (PUE and WUE)
Via Computed.Blg via Open Compute Project -----
Today (18.04.2013) Facebook launched two public dashboards that report continuous, near-real-time data for key efficiency metrics – specifically, PUE and WUE – for our data centers in Prineville, OR and Forest City, NC. These dashboards include both a granular look at the past 24 hours of data and a historical view of the past year’s values. In the historical view, trends within each data set and correlations between different metrics become visible. Once our data center in Luleå, Sweden, comes online, we’ll begin publishing for that site as well. We began sharing PUE for our Prineville data center at the end of Q2 2011 and released our first Prineville WUE in the summer of 2012. Now we’re pulling back the curtain to share some of the same information that our data center technicians view every day. We’ll continue updating our annualized averages as we have in the past, and you’ll be able to find them on the Prineville and Forest City dashboards, right below the real-time data. Why are we doing this? Well, we’re proud of our data center efficiency, and we think it’s important to demystify data centers and share more about what our operations really look like. Through the Open Compute Project (OCP), we’ve shared the building and hardware designs for our data centers. These dashboards are the natural next step, since they answer the question, “What really happens when those servers are installed and the power’s turned on?” Creating these dashboards wasn’t a straightforward task. Our data centers aren’t completed yet; we’re still in the process of building out suites and finalizing the parameters for our building managements systems. All our data centers are literally still construction sites, with new data halls coming online at different points throughout the year. Since we’ve created dashboards that visualize an environment with so many shifting variables, you’ll probably see some weird numbers from time to time. That’s OK. These dashboards are about surfacing raw data – and sometimes, raw data looks messy. But we believe in iteration, in getting projects out the door and improving them over time. So we welcome you behind the curtain, wonky numbers and all. As our data centers near completion and our load evens out, we expect these inevitable fluctuations to correspondingly decrease. We’re excited about sharing this data, and we encourage others to do the same. Working together with AREA 17, the company that designed these visualizations, we’ve decided to open-source the front-end code for these dashboards so that any organization interested in sharing PUE, WUE, temperature, and humidity at its data center sites can use these dashboards to get started. Sometime in the coming weeks we’ll publish the code on the Open Compute Project’s GitHub repository. All you have to do is connect your own CSV files to get started. And in the spirit of all other technologies shared via OCP, we encourage you to poke through the code and make updates to it. Do you have an idea to make these visuals even more compelling? Great! We encourage you to treat this as a starting point and use these dashboards to make everyone’s ability to share this data even more interesting and robust. Lyrica McTiernan is a program manager for Facebook’s sustainability team.
Personal comment:
The Open Compute Project is definitely an interesting one and the fact that it comes with open data about centers' consumption as well. Though, PUE and WUE should be questioned further to know if these are the right measures about the effectiveness of a data center.
Posted by Patrick Keller
in Architecture, Science & technology, Sustainability
at
07:52
Defined tags for this entry: architecture, data, energy, information, monitoring, opensource, science & technology, sustainability
Creating Energy from Noise Pollution
Via Architecture Source via Archinect -----
soundscraper. Source: eVolo 2013 Skyscraper Competition
Soundscrapers could soon turn urban noise pollution into usable energy to power cities. An honourable mention-winning entry in the 2013 eVolo Skyscraper Competition, dubbed Soundscraper, looked into ways to convert the ambient noise in urban centres into a renewable energy form. Noise pollution is currently a negative element of urban life but it could soon be valued and put to good use. Acoustic architecture, or design to minimise noise, has long been an important facet of the architecture industry, but design aimed at maximising and capturing noise for beneficial reasons is an untapped area with great potential. The Soundscraper concept is based around constructing the buildings near major highways and railroad junctions to capture noise vibrations and turn them into energy. The intensity and direction of urban noise dictates the vibrations captured by the building’s facade. Covering a wide array of frequencies, everyday noise from trains, cars, planes and pedestrians would be picked up by 84,000 electro-active lashes covering a Soundscraper’s light metallic frame. Armed with Parametric Frequency Increased Generators (sound sensors) on the lashes, the vibrations would then be converted to kinetic energy through an energy harvester.
The energy would be converted to electricity through transducer cells, at which point that power could be stored or sent to the grid for regular electricity usage. The Soundscraper team of Julien Bourgeois, Savinien de Pizzol, Olivier Colliez, Romain Grouselle and Cédric Dounval estimate that 150 megawatts of energy could be produced from one Soundscraper, meaning that a single tower could produce enough energy to fuel 10 percent of Los Angeles’ lighting needs. Constructing several 100-metre high Soundscrapers throughout a city near major motorways could help offset the electrical needs of the urban population. This form of renewable energy would also help lower the city’s CO2 emissions. The energy-producing towers could become city landmarks and give interstitial spaces an important function. The electricity needs of an entire city could be met solely by Soundscrapers if enough were constructed at appropriate locations, also helping to minimise the city’s carbon footprint.
Posted by Patrick Keller
in Architecture, Interaction design, Sustainability
at
07:17
Defined tags for this entry: architecture, energy, harvesting, interaction design, kinetic, monitoring, sound, speculation, sustainability
Saturday, April 20. 2013Why Computing Won't Be Limited By Moore's Law. EverVia ReadWrite via @chrstphggnrd
-----
In less than 20 years, experts predict, we will reach the physical limit of how much processing capability can be squeezed out of silicon-based processors in the heart of our computing devices. But a recent scientific finding that could completely change the way we build computing devices may simply allow engineers to sidestep any obstacles. The breakthrough from materials scientists at IBM Research doesn't sound like a big deal. In a nutshell, they claim to have figured out how to convert metal oxide materials, which act as natural insulators, to a conductive metallic state. Even better, the process is reversible. Shifting materials from insulator to conductor and back is not exactly new, according to Stuart Parkin, IBM Fellow at IBM Research. What is new is that these changes in state are stable even after you shut off the power flowing through the materials. And that's huge.
Power On… And On And On And On… When it comes to computing — mobile, desktop or server — all devices have one key problem: they're inefficient as hell with power. As users, we experience this every day with phone batteries dipping into the red, hot notebook computers burning our laps or noisily whirring PC fans grating our ears. System administrators and hardware architects in data centers are even more acutely aware of power inefficiency, since they run huge collections of machines that mainline electricity while generating tremendous amounts of heat (which in turn eats more power for the requisite cooling systems). Here's one basic reason for all the inefficiency: Silicon-based transistors must be powered all the time, and as current runs through these very tiny transistors inside a computer processor, some of it leaks. Both the active transistors and the leaking current generate heat — so much that without heat sinks, water lines or fans to cool them, processors would probably just melt. Enter the IBM researchers. Computers process information by switching transistors on or off, generating binary 1s and 0s. processing depends on manipulating two states of a transistor: off or on, 1s or 0s — all while the power is flowing. But suppose you could switch a transistor with just a microburst of electricity instead of supplying it constantly with current. The power savings would be enormous, and the heat generated, far, far lower. That's exactly what the IBM team says it can now accomplish with its state-changing metal oxides. This kind of ultra-low power use is similar to the way neurons in our own brains fire to make connections across synapses, Parkin explained. The human brain is more powerful than the processors we use today, he added, but "it uses a millionth of the power." The implications are clear. Assuming this technology can be refined and actually manufactured for use in processors and memory, it could form the basis of an entirely whole new class of electronic devices that would barely sip at power. Imagine a smartphone with that kind of technology. The screen, speakers and radios would still need power, but the processor and memory hardware would barely touch the battery.
Moore's Law? What Moore's Law? There's a lot more research ahead before this technology sees practical applications. Parkin explained that the fluid used to help achieve the steady state changes in these materials needs to be more efficiently delivered using nano-channels, which is what he and his fellow researchers will be focusing on next. Ultimately, this breakthrough is one among many that we have seen and will see in computing technology. Put in that perspective, it's hard to get that impressed. But stepping back a bit, it's clear that the so-called end of the road for processors due to physical limits is probably not as big a deal as one would think. True, silicon-based processing may see its time pass, but there are other technologies on the horizon that should take its place. Now all we have to do is think of a new name for Silicon Valley.
Personal comment:
Thanks Christophe for the link! Following my previous post and still in the context of our workshop at Tsinghua, this is an interesting news as well (even so if entirely at the research phase for now but that somehow contradict the previous one I made) that means that in a not so far future, data centers might not consume so much energy anymore and will not produce heat either. Friday, April 19. 2013Une start-up lance le chauffage par ordinateurVia Le Parisien via @chrstphggnrd -----
Une douce chaleur règne chez 4 MTec, à Montrouge. Ici, pas de chauffage central, mais une quinzaine de Q.rad, les « radiateurs numériques » de la start-up Qarnot Computing, hébergée dans les locaux de ce bureau d’études. A l’origine de cette invention, Paul Benoît, ingénieur polytechnicien aux faux airs de Harry Potter, croit beaucoup dans son système « économique et écologique ». Avec ses trois associés, dont le médiatique avocat Jérémie Assous, très investi dans la promotion du projet, une vingtaine de personnes planchent aujourd’hui sur sa mise en œuvre.
Related Links:Personal comment: While we are still located in Beijing for the moment (and therefore not publishing so much on the blog --sorry for that--) and while we are working with some students at Tsinghua around the idea of "inhabiting the data center" (transposed in a much smaller context, which becomes therefore in this case "inhabiting the servers' cabinet"), this news about servers becoming heaters is entirely related to our project and could certainly be used.
Posted by Patrick Keller
in Science & technology, Sustainability
at
05:31
Defined tags for this entry: climate, computing, data, energy, interferences, science & technology, sustainability
(Page 1 of 1, totaling 5 entries)
|
fabric | rblgThis blog is the survey website of fabric | ch - studio for architecture, interaction and research. We curate and reblog articles, researches, writings, exhibitions and projects that we notice and find interesting during our everyday practice and readings. Most articles concern the intertwined fields of architecture, territory, art, interaction design, thinking and science. From time to time, we also publish documentation about our own work and research, immersed among these related resources and inspirations. This website is used by fabric | ch as archive, references and resources. It is shared with all those interested in the same topics as we are, in the hope that they will also find valuable references and content in it.
QuicksearchCategoriesCalendarSyndicate This BlogArchivesBlog Administration |