December 1st Event

Thanks to everyone who came out on December 1st!

Here’s Tom Roberts presentation on the 2011 NEC Article 645:

IAEI Article 645 Presentation

For more information on the IAEI, please visit their website at http://www.iaei-michigan.org/

Posted in Uncategorized | Leave a comment

Dell Warranties Servers for ‘Fresh-Air Cooling’

http://www.pcworld.com/businesscenter/article/236801/dell_warranties_servers_for_freshair_cooling.html

Posted in Uncategorized | Leave a comment

Last Meeting PowerPoint

Hope everyone enjoyed the last meeting.  Here is the PowerPoint:  Conquering Chaos Cooling.

Posted in Uncategorized | Leave a comment

Inspiration for the Low Carbon Data Center

A great keynote by Robert F Kennedy Jr at the national 7×24 Conference.

http://www.greenm3.com/gdcblog/2011/6/13/inspiration-for-the-low-carbon-data-center-7×24-exchange-key.html?printerFriendly=true

Don’t forget about our chapters upcoming meeting on Tuesday!  Click here for more details.

Posted in Uncategorized | Leave a comment

Our Next Event

Sponsored by:

Chaotic Cooling

How to conquer cooling costs in your Data Center

Learn why most data centers are using methods that are best described as chaotic cooling

  • What chaos cooling is
  • How it occurs
  • The consequences on energy costs
  • How to resolve cooling issues
  • How to reduce energy usage
  • What low cost methods can be used

The presentation by Eaton’s Wright Line Business will be paired with a valuable end-user/owner presentation.

Audatex, a Solera Company, will discuss their recent efforts with a data center move and consolidation.  They will discuss their approaches with low cost/high reward efficiencies and standards.

  • When: Tuesday June 21
  • Where: Pi Restaurant (The Pub), 28775 Franklin Rd., Southfield MI 48034     (go to Banquet Hall behind Restaurant)
  • What: Light Dinner, Beer and Wine will be served during the event
Posted in Uncategorized | Leave a comment

Last Meeting Docs

Blue Cross Blue Shield Presentation

Raritan Presentation

Geist Presentation (coming soon)

Football Squares *WITH WINNERS*

Posted in Uncategorized | Leave a comment

Last Meeting – 1/27/11

Come learn about  the challenges of PUE and tools to help you better manage and monitor your PUE!

Meeting Highlights:

Data center power usage effectiveness (PUE), developed by The Green Grid, has received a lot of attention. But how do you use PUE to minimize wasted power? How do you identify stranded power and thereby possibly delay the construction of a very expensive new data center? How do you justify for management the investment to gather detailed power information?

Our meeting will start with Blue Cross Blue Shield of MI sharing with you why they did a PUE study, their results from the study and future goals for them to improve their PUE. Also we will have Geist IF and Raritan presenting on ideas of how to manage and monitor your PUE.

  • When: Thursday January 27 at 4:00 P.M. to 8:00 P.M.
  • Where: Pi Restaurant (The Pub), 28775 Franklin Rd., Southfield MI 48034     (go to Banquet Hall behind Restaurant)
  • What: Light Dinner, Beer and Wine will be served during the event

Click here for the meeting flyer.

Posted in Uncategorized | Leave a comment

EPO Vulnerability Fixed

Data center EPO vulnerability fixed in 2011 National Electrical Code
By Matt Stansberry, Executive Editor
05 Nov 2010 | SearchDataCenter.com
IT infrastructure news
Digg This Stumble Delicious Google Fusion
After years of wrangling with the National Fire Protection Association (NFPA) over the data center emergency power off (EPO) button, industry leaders have finally convinced the organization to modify its code to help companies avoid data center downtime.

The EPO, often a big, red pushbutton located near the data center exit, is designed to let firefighters shut down power quickly during a fire in the data center. Unfortunately, people push the EPO button accidentally or as an act of sabotage, causing what Uptime Institute founder Ken Brill has called “a corporate heart attack.”

A group of industry leaders from AFCOM and Uptime started working together on this project in 2007, finding contacts within the NFPA, researching the history of the EPO, and traveling around the country to build consensus and educate NFPA code panel members.

David Boston, data center consultant and former manager of the Site Uptime Network, said during the June 2010 meeting of NFPA members, a proposal to modify NFPA 70 Article 645-10 cleared its last opportunity for objections. It is now on its way to the Standards Board for final review and will be included in the 2011 National Electrical Code, due to be published late this year and issued in January 2011.

Boston said this will allow data center owners to construct or retrofit their facilities without needing to place an EPO switch at each principal exit. Instead:

“Owners will have the ability to negotiate a preferred EPO location within their facility with the local Authority Having Jurisdiction (AHJ), thereby reducing the risk of accidental or malicious power interruption. Owners who meet more stringent criteria will have the option, if the AHJ concurs, to construct their facility with no EPO (an alternate means of disconnecting power in the event of fire will be required).”

Data center EPO horror stories abound
Tom Roberts, an AFCOM board member and a data center facilities manager for Trinity Information Services, a healthcare company in Farmington Hills, Mich., has seen two data center outages due to an EPO misfire.

Roberts rolled out a brand new data center in the spring of 2003. Trinity had dedicated the building and was starting to bring clinical applications online when the facility took an EPO hit on Easter Sunday.

The event resulted in mass confusion and illustrated how vulnerable the systems were from a single source of interruption, Roberts said. And these buttons were scattered throughout the building.

That experience lit a fire under Roberts, who decided to take the issue to his local fire marshal, who agreed that the uptime of the healthcare provider’s data center was critical. The local fire authority allowed Roberts to leave the EPO system in bypass (meaning it wouldn’t shut down the systems unless the bypass was engaged by facilities staff). But due to a mistake in the building plan schematics, there was a loop around the EPO circuit that Roberts wasn’t aware of, so the bypass didn’t work. Trinity took another EPO hit in 2008.

At that point, Roberts vowed not to retire until this problem was addressed. “Now it’s fixed, I can’t retire because of the economy!” Roberts said.

Where did the data center EPO requirement come from?
Data center pros have battled the EPO requirement for years. Some huge companies threw their weight around with the NFPA, but to no avail. So how did this effort succeed?

Richard Schlosser, President at Baltimore-based data center design firm TiePoint Engineering, said getting the code changed depended on getting buy-in from the fire protection community, and understanding where the regulation came from in the first place.

Schlosser was on the committee to pursue the code change, and researched the history of the data center EPO. He spent a week in a library in Braintree, Mass., researching codes and figuring out when it first showed up in the National Electric Code, and traced it back to a huge data center fire in the Pentagon in1959.

According to the Arlington Fire Journal:

“The flames erupted in a computer room operated by an Air Force statistical agency and burned for more than five hours — fueled by magnetic tape. There were no sprinklers. The blaze scorched 4,000 square feet and caused $30 million damage to the massive building and the computer equipment.”

Schlosser said Congress then went to the NFPA and asked the group to write rules protecting data centers in the event of fires. “They came up with a whole new set of rules, including a kill switch for the data center console,” Schlosser said.

Scaling back safety requirements for fire service first responders wasn’t easy, Schlosser added. For the data center industry, putting a big red button next to the door that anybody can punch isn’t convenient. But fire service professionals consider the EPO a life-saving measure.

“We’ve created the environment [so] that the fire fighters think they needed a kill switch at the door. It was a big deal for them to relinquish that control,” Schlosser said. “They feel like it’s a big step backwards from a life safety perspective.”

Instead of fighting this uphill battle, the committee decided to try to befriend the authorities and provide some alternatives.

Schlosser said one option is to put a supervisor on duty 24 hours a day. This person would be trained and qualified to shut the place down and would be continuously available to meet fire responders when they show up and go through an approved procedure before entering the facility with axes and hoses. The other alternative is to move the EPO away from the exit doors into an area that is under control of management. “Firefighters can go into a locked room and shut down the data center from there,” Schlosser said. “It’s just like an elevator closet or electrical room. It’s not accessible to the main public in the building, but authorized people have access.”

Dealing with your local fire authority on data center EPO issues
While the new National Electrical Code is headed to the printer right now, and the change will be on the books for 2011, it could take one to three years for local jurisdictions to adopt the new code. For most data centers this will mean it’ll take effect in the 2013-2014 time frame.

What can you do to speed up the process? Roberts recommends asking your fire marshal, or local Authority Having Jurisdiction (AHJ) in fire-speak, if they will adopt the 2011 code early.

“If you’re a data center facility person, you need to get intimate with these guys,” Roberts said. “You need to know what their expectations are. Those guys have tough jobs, inspecting multistory buildings, old factories, but if I can help them learn the small area I know about, it’s better to both of us. If I can get close to those guys it helps.”

Posted in Uncategorized | Leave a comment

Cloud Computing and SaaS Migration

Cloud Computing and SaaS Migration
By Mike Klein on October 19, 2010

Software companies continue to evolve how they deliver software. Eighteen years ago when I started my first software company, our application was delivered on a stack of 3.5” floppy disks that needed to be inserted one disk at a time during the installation process. Over time, the software industry delivery model evolved to CDs and downloading the software over the Internet.

Today, many software companies are facing another transition in their delivery model. The Software-as-a-services (SaaS) delivery model, where the software is accessed through a browser or thin client and never installed on the user’s desktop, is becoming more prevalent across the industry.

SaaS has many advantages over the traditional software delivery model. The recurring revenue stream, simpler maintenance and application updates, and the lower cost of delivery and distribution are especially attractive for both the application provider and the end users.

However, unlike previous transitions which were changes in manufacturing or delivery technology, the transition to SaaS presents dramatic changes to both the migration and the delivery of the software solution.

Migration – I’ve lived through the headaches of rewriting a legacy code base from a 32 bit single-threaded application to 64 bit multi-threaded environment. Despite our programmers’ most conservative estimates, the project got bogged down in compatibility and migration issues. We eventually decided to cancel the project after spending over $1 million in the rewrite effort, because we saw no end in sight to the migration issues.

Many companies face a similar challenge of migrating from a desktop or client-server architecture to a multi-tenant SaaS model. In many cases, the transition to SaaS involves a re-write of the entire application, which can be fraught with delays and incompatibility with legacy systems. This can have a dramatic impact on time-to-market, especially when customers are pressing for a SaaS application in the short term with all of the same features they currently enjoy on their desktop or client-server solution.

Delivery – Another challenge for software companies migrating to a SaaS delivery model is the requirement to deliver the application online 24 x 7. At my last software company, we had 3 racks of servers in an ad-hoc data center (really just an office with an extra 10 ton air conditioning unit). With in-rack UPSs, we could survive short outages, but the loss of power or Internet connectivity, while impacting our development team’s productivity, didn’t impact our customer base. An outage would only inconvenience our customer’s ability to download the application. It would not impact when they could run the application.

SaaS changes the delivery requirements dramatically. Since the application is only available online, the solution needs to be available 24 hours a day without fail. For many software companies, this means that the IT delivery infrastructure suddenly becomes core to customer satisfaction. With SaaS hosting, the bar is raised from servers being mostly available, to always available and always online. Many SaaS companies consider colocation or managed server hosting so they can outsource the infrastructure expertise and focus on what they do best – software and application delivery.

Migration to SaaS with private cloud computing – A number of our SaaS customers at Online Tech have moved their client-server application to a private cloud computing platform as the first step in migrating to a SaaS delivery model. They leverage the benefits of a private cloud and virtual servers to deliver their solution over the Internet. The end user doesn’t know if they’re running on in a multi-tenant software application or on a virtual server dedicated to their instance of the software. They see the same set of features they had with the legacy system.

As opposed to the public cloud, the private cloud can deliver a secure platform where the end user can be assured that their data is safe and the network is secure. Each instance of the application can be spun up on its own virtual server within the private cloud. Every time a new customer is added, an additional virtual server is spun up, with the operating system, application and configuration preloaded and ready to go within minutes.

Conclusion – The private cloud can simplify the process of migrating client-server applications to SaaS. Once the hardware platform is in place, a new virtual server instance can be spun up for each new customer quickly and for minimal incremental costs. Migrating client-server solutions to the private cloud removes the risk of rewriting code to a multi-tenant application from the critical path and delivers the same end user experience as in the client-server model.

Lower Risk, faster migration and preserving the end user experience – all good reasons to consider private cloud computing in your migration to a SaaS delivery model.

http://resource.onlinetech.com/saas-cloud-computing/

Posted in Uncategorized | Leave a comment

Top 5 Reasons Why You Should Transition to the Private Cloud

http://www.onlinetech.com/resources/newsletters/august10/

Posted in Uncategorized | Leave a comment