Tonight I was reading on lights out data centers. Reducing overhead in the data center isn’t a bad idea but it can be taken to extremes.

The idea of lights out data centers means that all functions normally required in the data center can be performed remotely. This means that management and control and monitoring facilities can be shared among multiple data centers, saving money.

Hardware such as the HP/Compaq RilO card gives full remote console access to the servers. Many modern servers also have remote console capability via serial ports. We use these for emergency out of band control. These can certainly reduce the amount of time to respond to trouble situations.

On the bad side of lights out, there are many things that can happen that could be resolved with a fast staff response. As my previous posts reveal, I’m a great proponent of dealing with things long before they get bad enough to turn into a disaster. Cascading failures are another good example where quick response can head off and contain problems.

A good example of this is the trend of using absolutely minimum UPS runtime (15 seconds for flywheel) before starting a backup generator. Short run UPS power combined with the potential of a generator failing to start and long tech response time equals a great potential for disaster.

I think there are a lot more productive ways to reduce data center costs (virtualization, removing zombie servers, etc) than removing skilled hands that can resolve problems before they can impact customers.

Vern, SwiftWater Telecom

Every x86 server turned down is good for more than $400 per year in savings and it’s only going to get worse. Hard to believe many data centers aren’t pushing to make this happen. What would you do with an extra $400?

I see that Cisco has put a hardened router into orbit. I can just see the call out bill on this one when a buggy IOS release leaves it sucking sod.

Survey says, 41% of employees abscond with data. Do you have a plan for keeping your data and your customer’s data from walking out the door?

Vern, SwiftWater Telecom

La gamme Nexus est déclinée pour le blade center IBM qui peut ainsi bénéficier de l’apport du 10Gb et du FCoE

Voici quelques exemples de déploiements:

pour en savoir plus

Brian Doerr, CTO de SAVVIS décrit comment et pourquoi  un Service Provider se concentre sur des services d’infrastructure virtualisés pour aider les entreprises à réduire leurs coûts et à gagner en “agilité”. l’Unified Computing System de Cisco prend tout son sens dans cette stratégie

Tutor Perini est une société de construction américaine. A l’occasion de différentes acquisitions cette société possèdent 5 datacenters. Afin d’optimiser les coûts cette société a décidé de construire un nouveau datacenter partagé entre les différentes entités.

C’est dans ce cadre que s’est imposé le choix de la virtualisation VMware et  l’utilisation de l’UCS.

Pour en savoir plus.

L’objectif de ce document est de fournir aux lecteurs des informations sur les domaines d’application, les cas d’usage, les meilleures pratiques de la fonctionnalité Cisco Virtual Device Context, et ceci tout en respectant les principes de virtualisation logiques d’un même équipement physique.

Ce document est donc à destination des architectes réseaux, des ingénieurs, et des personnes opérationnelles ayant un intérêt pour le design, le déploiement, ainsi que les cas d’usage possibles de la fonctionnalité Virtual Device Context (VDC) disponible dans le système d’exploitation NX-OS qui équipe le commutateur « Datacenter class » Cisco Nexus 7000.

pour lire le document

Cisco WAAS : rappels

Cisco WAAS fait partie du portefeuille de produits de Cisco Application Networking Services (ANS), la plus aboutie et la plus vaste des solutions réseaux capable de comprendre les applications afin d’en améliorer leur valeur et leur utilisation.

Cisco Wide Area Application Services (WAAS) est une solution attractive d’accélération des applications et d’optimisation du WAN pour les sites distants, qui améliore les performances de toutes les applications basées sur TCP circulant sur un réseau étendu (WAN). Avec Cisco WAAS, les entreprises peuvent consolider les serveurs et éléments de stockage distants à l’intérieur de leurs Data Centers ou sites centraux, tout en continuant d’offrir un niveau de service similaire au LAN pour les utilisateurs des sites éloignés des centres informatiques.

Les solutions Cisco WAAS se présentent sous des formes différentes suivant les besoins:

• Une carte d’extension, Network Module, pour les routeurs ISR à partir du modèle 2811, idéal pour les sites distants en bordure du réseau WAN
• Une appliance adaptée à chaque topologie de site – Allant des besoins pour quelques utilisateurs jusqu’au Data Center, installée soit en coupure ou en redirection
• Un logiciel pour poste de travail, WAAS Mobile, pour les utilisateurs nomades

Nouvelles fonctions incluses dans la version 4.1.5a

En septembre dernier, la solution Cisco WAAS faisait également sa rentrée avec la nouvelle version 4.1.5a. Cette mise à jour met l’accent sur la maturité, la stabilité et les l’interface utilisateur. Les principales nouveautés sont:

• Intéropérabilité CIFS avec les nouvelles applications Microsoft : Windows 2008 R2, Windows 7, Windows Remote Administration Tools, accès à Microsoft Access avec Excel
• Intéropérabilité SSL : Windows 2008 avec IIS 7.0, Windows 2008 R2 avec IIS 7.5, SharePoint 2007 et SharePoint 2007/ISA
• Amélioration du Central Manager en haute disponibilité (Actif/Passif), possibilité d’administrer un plus grand nombre de boîtiers WAAS

pour aller plus loin 

Network Analysis Module (NAM) sur les plateformes Wave 574 et 674

La seconde nouveauté est la possibilité d’héberger localement  au sein d’une plateforme Cisco Waas une sonde logicielle de surveillance, d’aide au diagnostic et d’analyse du trafic réseau. Déjà proposé sur différentes plateformes Cisco comme dans les routeurs ISR ou dans les commutateurs Catalyst 6500 et routeurs 7600, la sonde Network Analysis Module (NAM) se décline pour la première fois sous la forme logicielle. En installant ce composant au sein d’un serveur virtuel sur les plateformes Wave 574 et 674, les clients peuvent disposer d’informations visuelles pour comprendre, gérer et améliorer les performances des applications délivrées aux bureaux régionaux et aux utilisateurs à distance.

L’intégration de la sonde NAM sur Waas  est  destiné aux réseaux disposant d’un Data Center de petite taille ou bien de sites centraux souhaitant une visibilité sur environ 4000 connections TCP optimisées. La sonde NAM est capable de mesurer les bénéfices apportés par les modules d’optimisation de trafic WAAS. La sonde collecte des statistiques sur chacun des modules WAE et offre des rapports montrant les gains de latence et de volume de trafic sur les différents segments du réseau entre un serveur et les utilisateurs. Dans un projet de déploiement de modules WAAS, la sonde NAM peut être utilisée avant le déploiement pour identifier le meilleur emplacement pour les boîtiers d’optimisation WAAS, et après le déploiement pour mesurer les gains apportés.

En installant la sonde NAM au sein d’un serveur virtuel Waas, vous bénéficiez de :

• Flexibilité et diminution du TCO en réduisant les extensions matérielles
  • Facilité d’installation – via Waas Central Manager
  • Pas besoin de matériel dédié – sonde logicielle uniquement sur le Waas central
• Une photographie complète des flux applicatifs et des usages du réseau en temps réel
• Un rapport mettant en avant les serveurs et les applications nécessitant une optimisation
• Une analyse présentant les résultats  avant et après l’installation de boîtiers WAAS dans un réseau

Une version d’évaluation du logiciel NAM de 60 jours est disponible pour tout client disposant déjà d’une plateforme WAVE 574 et WAVE 674

Pour aller plus loin 

[qi:gigaom_icon_hardware] As compute demand increases, demand for power in data centers is soaring. To help IT professionals halt the spread of watt-consuming servers, the industry needs to develop software that can communicate the ways in which the various layers of the data center perform and interact. They need a binary version of Cesar Millan — a data center whisperer.

Speaking at a panel held Wednesday night in Austin, Texas, several folks from the large server shops and a distinguished engineer who runs a data center for IBM (s ibm) spoke about the challenges of keeping power consumption down in a world where computing demand is going up. (For a truly in-depth look at this topic, check out our GigaOM Pro report — subscription required.) The panel went beyond just power and cooling (thank goodness) to focus on how companies are increasingly viewing power consumption in the data center as a whole, rather than merely as the sum of of the data center’s processors.

IBM’s Scott Winters said he saved 30 percent on his energy costs over three years while increasing his computing abilities by 50 percent and his storage by 150 percent. He did this in two primary ways: by virtualizing his data center and creating a pool of shared resources that are used on demand, and by paying attention to software he has running that tells him what’s happening on his servers.

“My data center was whispering secrets, and now I have a way to understand them,” Winters said. He said his IBM software and linking that software to the physical infrastructure helped him reach such an understanding, especially in regard to managing power consumption. It’s a strategy that HP has embraced with its products; there are also several startups pushing data center sensor networks that allow the data center’s server hardware and its physical infrastructure like the chillers and air conditioners to communicate.

But as the facilities and IT infrastructure merge (the jobs of the facilities manager and the IT manager are also on a path to merge, according to members of the panel) standards are needed. The folks building the physical infrastructure typically use proprietary software in their products and sensors and getting that sensor network to talk to your servers can require a big programming effort. Once folks can manage their physical infrastructure and their hardware, the next step is to tie the physical and hardware layers to the application layer. That’s a big dream, and we’re still far off. But given the demand for computing and constraints on providing the power to meet that demand, it’s an issue that panels like the one Wednesday night will help solve.

I’ve been watching this video of a hit on a server cabinet. It’s kind of like a drive by from the Data Center Mafia .

You knew it had to happen, there’s now a “Dummies” book for green data centers. I fear for the industry now.

Watch for my cloud data center series starting on Monday 11/23!

Vern, SwiftWater Telecom

TELEHOUSE America’s NYIIX Internet Peering Exchange Becomes the Largest and Longest Running Public Peering Point in the New York Market

Linked with TELEHOUSE America’s Los Angeles Exchange LAIIX, TELEHOUSE Peering Offers Service Providers a Global Interlink for IX Interconnection

Staten Island, NY, USA – November 18, 2009– TELEHOUSE America (www.telehouse.com), the United States’ leading provider of dedicated data centers, international Internet exchanges and managed IT services, announces that its New York International Internet Exchange public peering platform (NYIIX) remains the longest running and largest public Internet exchange point in the New York metropolitan market. Since its inception in 1996, the NYIIX has been a vital public IP interconnection point for global carriers, ISPs, content providers and enterprise businesses within New York. NYIIX is headquartered at 25 Broadway, TELEHOUSE America’s Broadway Center, and provides redundant and diverse direct interconnectivity to and from the 60 Hudson Street and 111 Eighth Avenue carrier hotel facilities.

The NYIIX platform continues to be updated to meet the growing technology requirements of the industry and currently operates Brocade RX series switches. The platform supports IPv4 and IPv6 peering connections and offers standard bilateral peering exchange as well as multilateral peering arrangements. In addition, NYIIX offers private peering over VLAN – enabling secure private peering or direct bandwidth over the NYIIX. As of October 2009, the NYIIX peering platform has over 120 members and handles over 80 Gbps of traffic, making it the largest public Internet exchange in the New York Metro Market.

In addition to NYIIX, TELEHOUSE America also operates the Los Angeles International Internet Exchange (LAIIX). Launched in 2000, the LAIIX provides public peering interconnectivity via its Layer-2 switch offering connectivity to multiple transit providers. In Los Angeles, the LAIIX provides optimal primary and secondary IP traffic routing options with direct route server access, ensuring highly advanced peering among multiple simultaneous network routes. As one of the oldest private peering points in the Los Angeles market, LAIIX serves as a key IP peering gateway to the Asia-Pacific region.

With the announcement of TELEHOUSE America’s Global Interlink services in 2008, the company interconnected its NYIIX and LAIIX platforms and has also made both peering exchanges accessible from any of the 12 Interlink PoPs throughout the world. Global Interlink provides seamless redundant and diverse Ethernet private line connections between multiple carrier hotels throughout the US and Europe.

“With over 120 active members on NYIIX and over 55 members on LAIIX, TELEHOUSE America provides robust public peering connectivity solutions among the most prominent network operators in the industry,” commented Akio Sugeno, Sr. Director Business Development, Internet Engineering & Operations for TELEHOUSE America and the founder of both NYIIX and LAIIX. “We are excited to witness the continued growth of our Internet Exchange platforms in both New York and Los Angeles and look forward to extending the NYIIX to 7 Teleport, TELEHOUSE’s Teleport Center on Staten Island in the coming months.”

Connectivity to the NYIIX and LAIIX is currently available from major carrier hotel facilities throughout the US and Europe, for more information about TELEHOUSE America’s public peering solutions and its complete total solutions for data center solutions, colocation, infrastructure management and global communications services, please visit www.telehouse.com or email sales@telehouse.com.

MEDIA CONTACT:

Ilissa Miller

Jaymie Scotto & Associates 1.866.695.3629

pr@jaymiescotto.com

TELEHOUSE AMERICA CONTACT:

Vincent Corley

Telehouse America

718.355.2572

corley@telehouse.com

This morning I was reading about evaluating data centers. What struck me about this article was how out of date some of the criteria people are applying is.

The first point is that a data center MUST have raised flooring and the prospective customer must know how deep it is. If you come to one of my data center facilities, you’ll be disappointed to see that there’s no raised flooring at all. All power and data cabling are carried overhead via cable racks and dropped to the cabinets. From an aesthetic standpoint, the cable racks are not as “neat” as the raised floor, since they expose the cabling. A large installation can also look a bit like a jungle gym (I’ve seen facilities with 6 or more levels of cable rack carrying different types of service). On the other hand, eliminating the raised floor reduces the unproductive load on the floor underneath, eliminates the need for underfloor fire detection and suppression, and makes cabling far easier (think about pulling dozens of tiles).

A simple walk through can certainly reveal a lot more about cooling problems than giving the 3rd degree about the exact equipment, cooling technique, and capacity. The hot and cold spots from lousy airflow control are all too easy to spot just by walking through them. It doesn’t matter if the cooling is legacy HVAC or modern green free air cooling, it matters that it’s done right and produces the expected results.

Total power and cooling capacity are irrelevant as long as the data center provider isn’t exceeding them. If there’s a known need for expansion capacity, then the discussion has to be had with the data center provider about reserving adequate capacity for expansion. Available power and cooling capacity can change radically from one day to the next in busy facilities. The important thing is that the provider knows exactly where they are on capacity. A provider who doesn’t have a handle on their cooling capacity is likely to blow right by it with the predictable catastrophe.

Insistance on the facility’s generators being internal to the building is questionable. While this increases security, it also places flammable fuels inside the building, not the best place. In my experience, most large facilities and most conversion facilities will have their generators in an outside secured area. Expectations of seeing a UPS in every cabinet or even massive whole facility UPS cabinets may be disappointed by modern green DC power plants.

These are just a few of my takes on the discussion. It’s important to remember that data center design is taking radical leaps now with the implementation of energy saving green techniques. If you go in expecting to see yesterday’s equipment in the data center, you’ll get yesterday’s data center.

Vern, SwiftWater Telecom

server co-location services

Nov. 20 – The company changed their sales estimation of 2009 to over 10%, because of, the demand to data center is growing in here. Regarding the world wide market estimation, it will be 25% growing, they said. Especially, in the financial, telecom and Web contents market, high growing will be expected.

J <http://itpro.nikkeibp.co.jp/article/NEWS/20091120/340821/?ST=system>

Tonight I’ve been reading about the FAA computer glitches that have plagued air travel. Reading this reveals that the FAA has raised the cascading failure to a new level.

Cascading failures have had a lot of play in data center and IT flops lately. The mighty powerhouse of Google had a major outage of its Gmail service because a maintenance gone wrong reduced router capacity below the amount needed for service, overloading other routers in a chain reaction. A number of major data center outages have been the result of cascading power failures.

In the FAA incident, a software misconfiguration during maintenance, leading to failure of a primary router and a backup router that failed to take the load. If you look at the previous outages, this is a common thread. Failure of both primary and backup, failure of a high reliability SONET ring, etc. As a former SONET engineer, I can say that nothing short of a major widespread disaster should disrupt a properly designed ring.

One of the quotes at the end of the article calls it out clearly. One part misbehaves and it disrupts everything else. It’s not the disasters like fire or earthquake destroying the facilities that should worry us the most, it’s the uncontrolled cascading failure.

So, how do we prevent the cascading failure? First, be prepared to recognize abberant results of maintenance procedures and have a way to back out of changes before they have major impact.

Second, avoid risky procedures of little benefit. This means no people with meters manually measuring voltage inside live power panels just to see if connections are good. Many of these failures have human error triggers.

Finally, make sure that all facilities have cutouts to prevent the spread of failures. You can’t always prevent mistakes but you can prevent a mistake from spreading into a catastrophe.

Make sure your data center or network isn’t next to be in the news for a preventable cascading failure.

Vern, SwiftWater Telecom

I just browsed across a post on evaporative cooling in Microsoft’s new data center containers. I’m not sure it’s the best idea to be humidifying the servers this way, and I’m darned sure it won’t be a good idea when something flips out and the humidity condenses inside this large steel box.

Best example of how NOT to do a planned outage. Burlington Coat Factory pulled the plug on it’s entire ecommerce site for 28 hours for hardware and software upgrades. No warning, no meaningful explanation during, just off the air. Just proves that anyone can serve as a bad example.

Coming up on Monday, I’ll be doing a post series on the engineering, construction, and launch of our new data center cloud!

Vern, SwiftWater Telecom

virtual private servers
virtual private workstations

Pour ceux qui envisagent le cloud computing , que ce soit sous la forme “private” ou “public” ,  la virtualisation est un passage obligé en partie pour supprimer l’adhérence application/harware . Cela permet ainsi à une charge applicative d’être “mobile” et donc capable d’utiliser une ressource de traitement disponible au sein d’un datacenter ou sur un autre datacenter.

 La mise en oeuvre de la mobilité des machines virtuelles à grande échelle met en évidence la nécessité pour chaque machine virtuelle de garder ses attributs (connectivité, sécurité etc..) lors de son déplacement. Le réseau ne doit plus s’arrêter aux machines physiques mais aller jusqu’au niveau des VMs.

Sur ce site VMware http://www.vmware.com/technical-resources/virtual-networking/ vous pourrez trouver beaucoup d’informations permettant d’adresser ces points.

Welcome to part 1 of my blog series, “Birth of a cloud” (and you thought these came from the cabbage patch!). In these posts, I’ll be taking you from the inception and planning of the data center cloud to an important mature component of the green data center.

For some time now, we’ve operated internal use and customer virtual machines using the excellent Xen hypervisor. As good as the standalone virtualization has been, it’s time to step it up a notch and take advantage of all the capabilities of a cloud, so we’re using the new Xen Cloud Platform as a good foundation.

For this project, we’re using Rackable Systems servers with 2x AMD Opteron 275 dual core CPUs, 8Gb memory, and 2GbE ports. These aren’t state of the art by any means, but they’ve been doing yeoman duty as standalone hosts and they’re rock solid hardware (we only use Linux virtual machines so there’s no need for hardware virtualization support).

We’re starting off with 6 cloud hosts but we’ve laid out the network for the full 16 cloud hosts that the Xen Cloud Platform we’re using is capable of (figuring 6 core Opterons, that’s 96 cores). The network uses a single switch, split in two networks, one for data traffic (all port 1s) and one for storage traffic (all port 2s).

Storage for the cloud is a common NFS server. This allows us to take advantage of XCPs workload balancer and the ability to shift live virtual machines between cloud hosts seamlessly.

The cloud is also integrated into the data center DC power distribution. Cloud hosts are split between separate DC power feeds which, with. the addition of XCPs work load balancer, will insure maximum reliability.

This is a basic overview of the hardware infrastructure to support a virtual machine cloud. In part 2, I’ll talk about establishing the cloud and bringing it up under manual control. In part 3, I’ll talk about using the work load balancer for automatic control of the cloud and finally, in part 4, I’ll talk about advanced projects that provide the ultimate in green results, such as the effort to allow the cloud to power down unneeded machines when demand is low. Stay tuned!

Vern, SwiftWater Telecom

First item up is the story that Virginia has had massive service failures in state departments because their $2.3B outsourcing deal with Northrop Grumman neglected to require network redundancy. What is it with providers these days just happily letting customers walk off the cliff? 12 outages and 100 hours of downtime in 5 WEEKS! It’s like IBM and Texas where IBM knew a piece of equipment was problematic and allowed it to fail without backup because backup wasn’t specified. Go the extra mile and be the hero, people!

Worst capacity planning award: Ebay kills its search functionality for most of the day Saturday. Gee, people are going to do more selling coming up to Christmas, who’d have thought?

Vern, SwiftWater Telecom

data center facility engineering
data center DC power plant engineering

Un petit outil sympa qui permet de visualiser un UCS dans son ensemble ou composant par composant. Il permet également d’en connaitre les mesures , bref d’en avoir un avant gout avant de linstaller dans son datacenter.

Visualiser l’UCS en mode interactif

Nov. 20- The company launched a new data center for the next generation computing in their famous Tatebayashi farm. It has 23,000  square meter of computing area, and provides robust security and low energy facilities. There are some of new infrastructure for the cloud computing, the company said.

Regarding PUE, there is no information.

J <http://www.rbbtoday.com/news/20091120/63906.html>

Tonight I was reading about green computing needing a data center whisperer. I believe there’s a lot more productive way to integrate the green data center and the cloud that lives in it.

Virtual server clouds certainly have a major impact on making the green data center greener still. Efficiency gains from running servers at much higher utilization rates by converting low utilized physical servers into virtual machines on a much smaller amount of physical servers is undeniable. Combine this with the ease of administration and maintenance of fewer physical boxes and the cloud looks like the greatest thing since sliced bread. The obvious reduction in power usage from fewer physical machines to support the same amount of workload is clearly a benefit to be data center’s bottom line as well as to reducing its carbon footprint.

The next trend that a lot of people are pursuing is fine monitoring of cooling in the data center using sensors that are already commonly integrated into servers. I’m never one to criticize gathering all available information, but in this case, I think this is a lot of unproductive effort. It’s possible to measure cooling with great amounts of granularity, controlling the cooling to that degree is next to impossible, at least right now. I think this is going to lead to large amounts of data overload for no result.

So, how do we take more advantage of the cloud and how does the cloud “live”? Let me give you a case example. We’re just completing commissioning of the core of our first virtual server data center cloud, based on the new Xen Cloud Platform. We chose XCP for a whole lot of fantastic features, ease of administration, and the ability to integrate it with our own front end, making it truly our cloud.

One of the features of XCP is the work load balancing. Create a new virtual machine and the work load balancer automatically chooses the physical server to put it on to even out the load across the cloud hosts. Shut down a cloud host for maintenance and XCP automatically shifts the virtual machines from that cloud host to a different one, without interrupting service to the customer.

Where these capabilities really shine is the potential to shut down physical servers completely during light load times and restart physical servers when the demand increases. This is the essence of a living cloud, it grows and shrinks to the need for it. Combine this automatic reduction of cooling in proportion and you have the ultimate power saving machine. Energy efficient servers, energy efficient cooling, all the capacity when you need it, and no underutilized servers and cooling sucking up power to no purpose. You get precisely the capacity you need when you need it, no more, no less.

This is the next evolution of the cloud’s contribution to data center efficiency. Is your cloud alive?

Join me starting Monday for my series “Birth of a cloud”!

Vern, SwiftWater Telecom

virtual private servers
virtual private workstations

Whether you’re at a dead stop (there are very few of you out there) or 15 percent there (like most of you) enterprises appear to be at a virtualization plateau and driving to the next level of relies more on an IT organization’s ability to evolve their people and processes versus deciding on the technology.

Don’t get me wrong, picking the right technology is extremely important–just ask any manufacturer! And much of the technology, like virtualizing desktops,  has matured and CIOs are more confident to deploy.  But not having the skilled people and honed processes are, in my estimation, an order of magnitude more important.

According to HP’s research, the biggest barriers or inhibitors to virtualization are: upfront costs, staff training/experience, gaining approval from corporate management, and infrastructure planning.

At World Wide we recognized the importance of people & processes, so we offer a variety of services including virtualization consulting that assesses an organization’s “operational readiness.” It includes an examination of the “as is”  infrastructure as well as the management tools and processes.   Steve Chambers, a Cisco virtualization evangelist in the U.K,  says that Cisco has put this “readiness assessment” in a framework called VOMA (Virtual Operations Management Assessment).  VOMA is about “institutionalization of technology across six dimensions:”

1. Management Processes
2. Organization
3. Governance
4. Technology
5. People
6. Metrics

via ViewYonder » Poor IT hygiene is a barrier to technology adoption.

The field of players is expanding and CIOs now benefit from more virtualization  infrastructure choices than ever.  Experienced CIOs know, while the technology is important, capitalizing on the benefits of virtualization is more about the people and the process.

I had an earlier post, Stand Up and Get Ready, that was more about manufacturers and resellers being ready for the economy to turn.   I now ask CIOs and IT leaders as a whole….

Are you virtualization ready?  More importantly, are your people and processes ready?