BIoT Canada

A Conversation With: John Siemon

Siemon's Executive VP of Engineering buoyed by the fact enterprises are 'outgrowing their cabling plants, their data centres and their IT infrastructure in general.'

March 1, 2008  

Print this page

CNS: Let’s start with the economy. How bad is it out there and do you see any silver linings for Siemon and the structured cabling industry as a whole over the next 12-18 months?

Siemon: Obviously, there is quite a bit of uncertainty in the global economy as a whole and it would be a bit foolish to think that the cabling industry is completely immune. That said, we see growth projections for the enterprise cabling industry that outpace overall economic growth. There is good upside potential for 2008.

The concerns of the larger economy are tempered by the fact that enterprises are outgrowing their cabling plants, their data centres and their IT infrastructure in general. Many of the build-outs that were completed in anticipation of Y2K are overdue for replacement or a major overhaul. From our perspective, the demand for cabling is still strong and growing.

CNS: You attended the recent BICSI Winter Conference in Orlando. What impressions did you come away with?

Siemon: Although there were no show-stopping product announcements, the sessions showed that technological advances and the innovation drive have not slowed. The presentations I saw point to evolving needs in the industry and networking in general. For example, there was an interesting parallel drawn between IT networking and centralization of the power grid in the early 1900s. If that analogy holds, the importance of IT infrastructure will only increase. While requirements for increased storage and transmission capacity will continue, the opportunity for innovation will more likely come from other challenges.

Manufacturers can tell you what’s coming from a product standpoint, but you can better judge the mood of the industry by talking to the attendees. If I were going to play Alan Greenspan, I would define the BICSI mood as “cautious optimism” with a good deal more optimism than caution.

More specifically, I noted a continuation of a trend towards the acceptance of screened and shielded systems. If you went to the 2004 BICSI and predicted this level of interest in screened and shielded cabling in North America you would have received some skepticism.

Siemon did just that and, yes, there were skeptics. But with each year, acceptance has grown. This year, I would characterize the acceptance of screened and shielded cabling as the best 10GBASE-T option as all but universal.

There is too much evidence on the potential benefits of F/UTP and S/FTP cabling for people to ignore it anymore. I would credit the few screened and shielded “pioneer” companies for that shift.

CNS: Our cover story this issue focuses on the data centre. How important are standards and what impact would you say TIA-942 has had in this space since it was first introduced in 2005?

Siemon: Standards provide minimum requirements for an infrastructure that can accommodate diverse needs and ongoing changes to the networking environment.

One of the most significant aspects of TIA-942 is the direction to use a structured cabling system and build on the ‘568 series of standards.

In the past, directly connected equipment has lead to a significant increase in abandoned cable, which is an issue in a majority of older data centres today. There are still examples of equipment rooms and data centres that are lacking in even the most fundamental principles of infrastructure management. While these images help to liven up a presentation or article, the IT facilities they depict are real. If those types of issues are not addressed, the risks associated with erosion of data centre infrastructure to both up-time and operating costs increase exponentially.

The data centre standard assists IT professionals in the design and installation of cabling that will support multiple generations of equipment without the need to re-cable. Realistically, it is far less disruptive to support network upgrades that are minimally invasive to the data center environment and do not require new cabling or cabinets.

Work is also progressing on ISO/IEC 24764, the draft international standard entitled “Generic cabling systems for Data Centres.” It uses ISO/IEC 11801 as the foundation for specifying a modified structure and configuration for generic cabling within data centres used to support existing and emerging applications. The implementation options cover connection schemes that reflect a wide range of operating environments. The votes on the second and hopefully final committee draft ballot are due in April.

CNS: Siemon recently introduced a 10G 6A F/UTP version of its MapIT intelligent infrastructure management system. Is it safe to say that 10G’s time has arrived?

Siemon: Absolutely. My answer would have been the same in 1999, but I would have been referring to a different phase in the 10Gb/s lifecycle. Back then, the time had come to start developing a 10Gb/s standard.

When pre-standard Category 6A solutions hit the market, the commercial availability of 10Gb/s product signified that the time had come. More recently, ratification of the 10GBASET standard had the same kind of effect.

With approval of Category 6A/Class EA standards, the installed base will continue to grow, which will lead to future generations of 10Gb/s electronics that are more readily available and affordable.

The current phase of the 10Gb/s life cycle is significant in that the “early adopter” phase is behind us. In our experience, the vast majority of mid to large enterprise projects include 10Gb/s capability in their initial specifications, and the number installing these systems is growing rapidly.

I believe that this trend is aided by the variety of 10Gb/s options and the resulting ability to choose a best-fit solution. North America in particular is beyond the limited de facto options of UTP and fiber. The market acceptance and manufacturer support of screened and shielded systems has changed the landscape considerably.

CNS: At the CNS 10th anniversary panel held in November you suggested that standards for the cabling infrastructure should get out in front of the applications and user needs. How easy or difficult is it going to be to get to that point?

Siemon: I do not believe it will ever be easy to gain consensus within standards groups made up primarily of competitors.

Having served on these bodies for so long, I can tell you that most participants truly want what is best for the users and the industry as a whole. Agreeing on what is best is a different story.

But, we know it can be done. Category 7/class F standards were published by ISO/IEC in 2002, four years before the ratification of the IEEE 802.3an 10Gb/s application standard. It was used to prove technical feasibility before formation of the task force and was the first and only standard-based cabling to support full 100m implementations of 10GBASE-T until 2008.

More recently, an amendment to ISO/IEC 11801 created the pending class FA, which is targeted to support the next generation of data applications beyond 10GBASE-T.

An interface standard has already published in support of class FA in the form of the second edition IEC 61076-3-104 standard. We know we can get out ahead of the applications.

On the other hand, we did not have approved TIA or ISO standard for Category 6A/Class EA channels until February 2008 — two months after IEEE moved ahead with projects to develop standards for 40Mb/a and 100Mb/s transmission.

I believe we can attribute this timing to the challenges faced pushing 10Gb/s through UTP cable and the RJ45-style interface rather than an inherent deficiency in the cabling standards process. Regardless of the underlying cause, the fact remains that this delay has put mainstream twisted-pair cabling behind the applications. What’s more troubling is the uncertainty surrounding what will come after Ca
tegory 6A. Fiber on the other hand is already well positioned to support 100Gb/s transmission.

CNS: You also talked of a technological shift that could soon see the “killer apps” being replaced by “app killers.” What is an app killer in your mind?

Siemon: By app killer, I mean the type of change that forces us to rethink fundamental aspects of network performance.

It is clear to me that trends in network utilization, operating cost and environmental considerations will take us in new directions that bring latency and power to the front of the line as the primary drivers for future networking technologies.

The ratified 10GBASE-T standard enables availability of transceivers that are commercially viable and will work with “generic” cabling, but it came with tradeoffs to cost, power and latency.

Unless those limitations are effectively resolved by 2nd and 3rd generation transceivers, commercial acceptance will never reach the levels of 1000BASE-T. Until then, other media options like fiber and other networking protocols like Infiniband will be called upon when power or latency are considered to be critical.

The good news is that technical feasibility of copper-based solutions for 40G and 100G data rates has been proven although it remains to be seen what form those solutions will take.

I am confident that there will be at least one standards-based copper cabling solution that will support transmission rates in excess of 10G with less power, lower latency than today’s 10GBASE-T. To do so, other tradeoffs may be necessary regarding interface type, cabling implementations and the ability to auto-negotiate to lower transmission rates.

CNS: What is the best way for an organization to maximize its cabling investment?

Siemon: Plan ahead and future proof. Provided that you intend to be in the facility for longer than five years, install the highest performing cabling available to maximize it’s lifecycle. The longer the cabling plant can support the network, the lower the total cost of ownership.

As an example, the initial installed cost of a Category 5e system is about a third the cost of a Category 7A installation. But, the Category 5e system can only expected to last 5 years. Category 7A, with performance beyond 10Gb/s can be conservatively predicted to last 15 years.

If you annualize the Category 5e and Category 7A (class FA) installation costs by their respective 5 and 15-year lifecycles, they are roughly equivalent. Add in costs associated with migration to Gigabit and 10Gb/s speeds, such as testing, remediation, removal and associated downtime, and the costs are reversed.

Category 7A annualized net costs are projected to be about one third the cost of a Category 5e solution without even factoring the cost of the new cabling to replace it.

CNS: Siemon certainly has an international presence. How important are your non-U. S. markets?

Siemon: Our ability to participate as a global market provider is absolutely critical. Siemon has invested heavily in regionalizing our global operations and support infrastructure.

We maintain major regional operations and support headquarters in North America, EMEA, APAC and CASA as well as regional Siemon sales offices in about 30 countries, and of course, provinces.

CNS: What are your plans for Canada moving forward?

Siemon: Although Canada was our first international market and we have been doing business here for nearly 50 years, our approach is consistent with the rest of the world. We continue to build an infrastructure and business relationships that deliver excellent service and the same high quality and performance of our installed cabling systems anywhere in the world.

Like any major global region, we view Canada as a unique market with specific needs and conditions. We want to continually learn and address those needs with in-country resources and infrastructure.

It is about establishing a market presence that is supported by strong channel relationships with distributors, installers, consultants and system integrators that are experts and positioned to provide service, innovation and value to our customers in Canada.

CNS: Finally, the IEEE 802.3 High Speed Study Group has been in operation since July 2006, ostensibly to bring 100 Gb/s Ethernet into the mainstream. When can we expect that to occur and what will drive the need?

Siemon: The project was approved by IEEE last December with a plan calls for standard approval by mid 2010. Beyond that, I hesitate to make predictions on when we will see 40Gb/s or 100Gb/s products hit the market.

I believe that 100Gb/s will be driven initially by the need for faster data center links to support continued increases in network traffic, particularly as it becomes more content-rich and bandwidth-hungry.