Next Gen Data Center Forum Circa 2004-2006



With the data center firmly established as the information hub of the modern business, it is vital that IT managers and users keep up on the latest news in technological research and networking developments. The NDCF met this crucial need with its in-depth news and analysis, product reviews, discussion forums, newsletters, and training programs.

This was the NDCF's website for a number of years until it merged with two other sites to become Byte and Switch.
Content is from the site's 2004-2005 archived pages offering just a glimpse of what this site offered its visitors.

About NDCF

The Next-gen Data Center Forum is a Website designed specifically for the needs of those IT managers who are responsible for data centers. But the NDCF is much more than just a source for news, research, and expert advice. It is also a Forum in the truest sense of the word, an online meeting place where IT managers can discuss the latest data center issues. This is where IT managers come to share data center knowledge and best practice.

 

2004


Data Center Infrastructure Concerns 06.21.04

Q. Does anyone have any concerns on the infrastructure needs of the data center? Specifically, cabinet and rack requirements? Are there good solutions available? Or do changes need to be made for this type of product?


A. In terms of infrastructure concerns, that is about the highest priority on most data center managers' agendas today- particularly cooling and power. With the proliferation of smaller and smaller equipment, and the surging use of storage racks, placing a lot of heat-producing devices in a very small footprint and close proximity [means that] cooling and power problems abound. As do the solutions being offered.

We have a full track on this at our upcoming [October] conference in Atlanta. We call the track "Facilities." This track focuses on such things as cooling, power, fire suppression, physical security, data center design, and raised flooring.
- Leonard Eckhaus, Chairman and CEO, data center managers' organization Afcom

 

Handling an FTP Session During a Failover 03.18.04
Q. How do you handle an FTP session in progress during a link failover to maintain its connected sockets?

A. The art of building reliable network and computing infrastructures is deep and complex. Generally, networks are architected to contain multiple redundant paths between any data source and any data sink. Those redundant paths may exist at multiple layers of the OSI protocol stack, with each implementation choice protecting against a certain class of failures and introducing its own set of complexities.

In your question, you ask specifically about failure of a network link. There are many ways to protect against this type of failure. Between two different network devices (switches or routers, for instance), there may be Layer 1 alternatives such as link aggregation, Layer 2 alternatives such as spanning tree (with or without the various rapid reconvergence variants), or Layer 3 alternatives such as various routing protocols (OSPF, BGP, etc.). Between a host and the network, options typically include Layer 1 options such as link aggregation or Layer 3 options such as providing a host with dual network interfaces.

Each of these various alternatives is agnostic with respect to application-layer protocols such as FTP; they work equally well with HTTP, SMTP, etc.
- Dave Roberts, VP of Marketing and Strategy, Inkra Networks

 

Backup Bottleneck 03.16.04
Q. What do you see as being the next major bottleneck in the data center?

A. As far as the next major bottleneck in the data center is concerned, my view is that the volume of data that needs to be backed up for disaster recovery purposes is growing faster than the bandwidth that is economically available to enterprises to move off-site
- Geoff Bennett, Chief Technologist, Heavy Reading

 

Upcoming Webinars

Date: 8/12/2004 2:00 p.m. New York/ 7:00 p.m. London time
Event: WiMax Metropolitan Area Wireless Networking Technology
Sponsors: Alvarion and Proxim
Speakers: Justin Springham, Senior Editor, Unstrung
Click Here to Register
WiMax is the hot new technology that has got carriers, component vendors, and equipment suppliers interested in metropolitan-area wireless coverage all over again.

Date: 8/17/2004 2:00 p.m. New York/ 7:00 p.m. London time
Event: RPR: Ready for Prime Time
Sponsors: Cisco, Corrigent and Luminous
Speakers: Scott Clavenna, Chief Analyst, Heavy Reading
Click Here to Register
Resilient Packet Ring (RPR) technology is reaching critical mass driven by demand from major incumbent service providers in every corner of the world. This event, timed to coincide with the standardization of RPR by the IEEE, evaluates what's on offer and looks at who stands to gain.

Date: 8/18/2004 2:00 p.m. New York/ 7:00 p.m. London time
Event: Scaleability & Reliability of Routing Protocols and OS in Public Networks
Sponsors: Data Connection and Nexthop
Speakers: Craig Matsumoto, Senior Editor, Light Reading
Click Here to Register
The software side of routers has always been important. Now it's becoming even more so, for two reasons. First, the emergence of new services, such as MPLS-based VPNs, is putting routing protocols like BGP under a lot of pressure. Second, scaleability and reliability (including the ability to execute hitless software upgrades) has become a critical for carriers moving to converged, packet-based backbones. This webinar identifies issues and reviews how vendors are addressing them.

Date: 8/19/2004 2:00 p.m. New York/ 7:00 p.m. London time
Event: Next Generation Services: Management Matters
Sponsors: Cisco, Micromuse and Smarts
Speakers: Peter Heywood, Founding Editor, Light Reading
Click Here to Register
The telecom industry is going through a fundamental change from focusing on building the next-generation network (NGN) to working out how to use that infrastructure to enable next-generation services (NGS). This marks a dramatic shift away from thinking of services in terms of network connectivity and towards thinking of services as applications, or bundles of applications, provided over whatever access infrastructure is most appropriate at the time. We analyze this critical trend.

Date: 9/1/2004 2:00 p.m. New York/ 7:00 p.m. London time
Event: ATCA: The Architecture of Tomorrow's Telecom Systems
Sponsors: AMCC, BATM Telco Systems, Force Computers & Intel
Speakers: Simon Stanley, Founder & Principal Consultant, Earlswood Marketing
Click Here to Register
The introduction of AdvancedTCA components and systems creates a new and significant opportunity for semiconductor vendors, chassis and backplane vendors, telecom blade manufacturers, and networking system vendors to expand their businesses and bring down system costs.

 

2005


Blade Center Solutions 03.14.05
Question: I have a blade center and a blade center cabinet. We plan to populate the cabinet with six blade centers -- using every rackmount unit. I'm looking for a solution that allows me to have my horizontal crossconnect (fiber and copper) installed in the cabinet without using rack space (there is non left). I need something that is on the back of the cabinet, but sticks out away from the blade centers. I don't want to run patch cables from each blade center to the network and SAN. Is this possible? I've looked everywhere.

Answer: I contacted our partner Dave Kohler, CEO of LOOK Communications in Columbus, Ohio, and he tells me the following:

There are two ways to accomplish this. First, is to have deeper cabinets that have front and rear rails. The depth makes it so that the infrastructure does not interfere with the server management hardware. The infrastructure is then placed on the rear rails at the top of the cabinet. The increased depth should be approximately 12 inches for the patch panels to patch properly.
Second, if the cabinets are already procured or existing, the use of raised floor enclosures enables connectivity. These enclosures are placed in the access floor at the rear of the server cabinet. Because of the high density of the enclosure, we generally utilize one enclosure per two server cabinets. A piece of innerduct approximately 2 to 3 feet long is placed from the enclosure to the server cabinet bottom rear. Optical fiber or category copper cabling with a properly listed jacket is then installed from the network and SAN switch edges to the enclosure. This does away with the use of long and custom patch cables. All connections are then accessible at the server cabinet location. The enclosures come in different depths depending on the raise of the access floor. For example, an 8-inch access floor enclosure can accommodate 192 optical fiber and/or copper connections. The cable media can be mixed in the enclosure if the network is category copper and the SAN is optical fiber.


- Marlin Ness, Practice Director of Enterprise Management, Greenwich Technology Partners

 

Management Structure 03.14.05
Question: I'm looking for best-practice information with regard to the organizational management structure of a large data center (120,000 square feet). Any data describing roles, responsibility, and reporting relationships would be appreciated.

Answer: This can best be answered by understanding the IT products and services that will be provided by the data center. There are several feasible operating models and a few questions should be addressed to give you the practical answer. For example:

What functional areas and systems will be installed in the data center?
Will the silos include telephony, mainframe, LINUX, UNIX, Windows, NAS, SAN, networking, security, middleware, and facilities infrastructure, etc?
Are you following a functional silo model or are you providing products and services in accordance with the IT Infrastructure Library (ITIL) IT Service Management framework?
How does the rest of the organization support the data center? Will the data center provide 24-hour or lights-out support?
What is the data center's relationship to other support areas?
The answers to these questions will provide the basis for your solution.

Greenwich Technology Partners is a definite proponent of ensuring end-to-end services for the business and recommends alignment in that manner. Your organizational structure and total staffing will depend on the skill sets of your personnel and their abilities in multiple functional areas. There are many reasonable options and they could probably all work.

However, you should ensure that there is strong documentation of roles, responsibilities, processes, procedures, and working instructions. You should also ensure clear lines of operational direction and control. Reasonable job descriptions should be tied directly to the responsibilities and the purview for each position. I would check the Web for plenty of examples.

Pet Management
Question: This is in a lighter vein, but it's a real question. The owner of a mid size data center brings his dog to work. The bare metal techs say that's a no no. His employees agree, but it's the boss? Suggestions?

Answer: The only way that's ok is if facility has positive air pressure (so always flowing out). But even then the dog must always be downwind of the servers. Ask the boss if he accepts the risk. RJ Daniels has exactly this set up and his dogs (plural) even have their own space with their large round dog beds and feeding station. Visitor often mistake the dog beds for floor pillows because of the designer fabrics and their size, but when they see the dogs, they crack up (chihuahuas). RJ says he's never had a problem with the air quality in the server room, but he thinks the risk is worth the quality company of his dogs. IMO, no dogs in the server room.

- Scott Miller, Senior Network Engineer, Greenwich Technology Partners

 

Fiber or Copper?
Q. Are fiber optics the standard media in data centers or does copper still have a large market share?

A. Both copper and optical share data center market share. I would guess that copper is still more dominant...
- Jon Oltsik, senior analyst, information security, Enterprise Strategy Group

 

2006

In 2006 the Next-Gen Data Center Forum was combined with Storage Pipeline and Byte and Switch.


Welcome to Byte and Switch

Byte and Switch welcomes readers of and Next-Generation Data Center Forum. We think you’ll find the combination of all three sites an unprecedented forum for storage networking professionals.

Even if you've visited Byte and Switch before, take a good look around. We’ve refurbished our site, making it easier to get to the content you care most about—news and opinion.

We’re redoubling our efforts to write for the enterprise IT buyers of storage—how you make procurement decisions, deal with management issues, and work with suppliers to drive innovation.

If you’re a new Byte and Switch reader or returning, you can also count on us for meaty content that helps you get to grips with a particularly thorny technical issue. Whether it’s SAN versus NAS, Fibre Channel or iSCSI, or a toss-up between Continuous Data Protection (CDP) and Data Protection Manager (DPM), our experts and end-user case studies can help you make more informed decisions.

What’s more, Byte and Switch continues its focus on original reporting on breaking news stories. In addition, we will provide ongoing information on the industry, helping you track storage-related stocks and keep current with a ticker of headlines about the public companies making news on a daily basis.

We have taken aboard the editors of all publications. In addition to the existing team, we are proud to announce that Terry Sweeney, who originated Storage Pipeline, is now Editor in Chief of Byte and Switch. James Rogers, formerly Site Editor of NDCF, has joined as well.

We're also happy to say that long-time Storage Pipeline columnist Jon William Toigo is making the transition with us to Byte and Switch. Look for his provocative insights and occasional rants twice a month. We've also archived on Byte and Switch the last year's worth of his columns from Storage Pipeline.

We're pleased to be partnering with CMP's Network Computing for product reviews and lab tests that really bang on those arrays and put storage software suites through their paces.

Like what you see? To get the full benefit of Byte and Switch, we encourage you to click here to become a registered user. This gives you full access to our message boards, newsletters, industry directories and specially targeted Webinars.

Think we missed something? Or maybe you have a hot news tip or cool case study to share with us. We welcome the dialogue—drop us a line with your thoughts and let us know what you think of the storage industry's largest and newly refreshed web site.

 

Next-GenDataCenterForum.com