Uptime vs. TIA-942: Introduction, why this series of articles?

Published on May 8, 2017

Edward van Leent
Chairman & CEO at EPI Group of Companies

Article 1 | Uptime vs. TIA-942: Introduction, why this series of articles?

During a recent one month tour throughout the USA and Asia I had the pleasure to meet numerous data centre owners/operators, consultants and end-users to talk about data centre trends and the challenges they are facing. During those conversations, we also discussed quality benchmarks for data centres facilities including the various standards and guidelines.

I started spotting a clear trend that there is a lot of misperception about data centre facilities benchmarking in relation to ANSI/TIA-942 vs. Uptime. Some of those misperceptions are based on outdated information as some customers didn’t keep up with the developments in that space as well as deception, created by some parties not representing the facts truthfully either by ignorance or intentionally for commercial reasons.

It was clear to me that the market needs to be updated on what is happening, this including the true facts of the matter. That’s what brought me to the idea of writing a few articles about this subject matter to ensure the market gets appropriate, fact based and updated information. I will address in a series of articles a variety of aspects and I hope that this will contribute to a more clear and fact based picture of the current situation and it will hopefully answer any question you might have regarding this subject matter. If you have any suggestions in terms of topics to be covered then please feel free to drop me a note at; edward@epi-ap.com.

Article 2 | Uptime vs. TIA-942: A short history

Before getting into the details of Uptime vs. TIA-942, I thought it would be a good idea to provide a bit of background so that some of the matters that will be discussed in upcoming articles can be seen in the light of the bigger scheme of things.

Uptime (UTI) came up with the data centre classification scheme based on four (4) different levels which probably all readers of this article know are indicated by the term “Tier”. It was first released in 1995 with the title “Tier Classifications Define Site Infrastructure Performance”. In 2005, this title was update to “Tier Standard Topology” also referred to as TST.

In the early 2000’s the TR42 committee of TIA decided to create a telecommunication standard for data centres. UTI and TIA got in touch with each other and UTI gave TIA the legal right to use the Tier philosophy it had developed for inclusion into what ultimately became the ANSI/TIA-942 (TIA-942) standard. There were a few key differences such as that TIA-942 did not only address Electrical and Mechanical as defined at a high level in the TST, but also included many other factors in two additional sections being Architectural and Telecommunication (I will expand more on some of the key (technical) differences in another article). Both UTI and TIA were using the term Tier to indicate the four different levels of design principles. UTI was, and still is, using the Roman numerals (I, II, III, IV) whereas TIA was using the Arabic-Indic numerals (1,2,3,4).

TIA released the ANSI/TIA-942 standard in 2005. The standard very quickly became popular for a variety of reasons. This was amplified when a number of organizations started to perform conformity assessments based on the ANSI/TIA-942 which clearly was creating a much more competitive environment in the market place where previously UTI was pretty much the sole player. There was also some level of confusion in the market when organizations were talking about having a Tier-X data centre without providing the reference as to whether this claim was based on either UTI-TST or the ANSI/TIA-942. These reasons slowly became more and more of an irritation point and in 2013 UTI approached TIA with the request for TIA to drop the term ‘Tier’ from the ANSI/TIA-942 standard.

TIA, being a non-profit organization, had no issues with that and as such it was mutually agreed upon that TIA would strike of the term ‘Tier’ from the ANSI/TIA-942 standard and replace it with the term Rated/Rating in the 2014 version of the Standard. In an upcoming article, I will discuss in more detail about the rights of using the term Tier and/or Rated/Rating as there are unfortunately some misperceptions about the legal rights with respect to the usage of the term ‘Tier’.

The above episode basically ended the relation between UTI and TIA and each of the parties are now working individually on the current and future versions of their own independent documents.


Article 3 | Uptime vs. TIA-942: Standard or guideline?

There have been many debates on the internet to discuss this topic including the confusion about its relation to codes and arguments about using a capital letter to indicate the term Standard. I think it is good to go back to one of the first definitions (as far back as 1667) which defined a Standard as ‘a specified principle, example or measure used for comparison to a level of quality or attainment’.  A guideline was defined as ‘A non-specific rule or principle that provides direction to action, behaviour or outcome’.  These definitions of course still leave some level of interpretation about what exactly can be identified even to the point that some would argue that the both terms can be used for the very same thing. I would argue that a Standard has a few important factors;

  1. Standards are developed by an accredited SDO (Standard Development Organization). This title is awarded by any of the three key members of the WSC (World Standards Cooperation) or their regional or national members who have been given the authority to accredit SDO’s. At a regional level you would have for example CEN which is the European standards body issuing EN standards. At the country level you have for example ANSI in the USA, BSI for the UK, SPRING in Singapore etc. Virtually any country in the world has their own.
  2. The development of the Standard is following a transparent development process as laid down by the organization which is governing the SDO development efforts. This typically includes key points such as that the process should be documented and available for others, members involved should be balanced etc.
  3. SDO’s are typically non-profit organizations
  4. SDO’s do not perform audits nor do they provide certification
  5. All requirements of the standard are transparent i.e. ALL requirements are available to those who wish to have insight in the standard and, just before you even ask the question; NO, this does not mean that the standard should be available for free.
  6. The Standard must be reviewed on a regular basis not to exceed 5-years. The outcome of that review will yield in either one of the three options, reaffirm, revise, withdraw.
  7. The intellectual property (IP) extends only to the standard itself and not to its use. This means that others than the SDO can use the material for various purposes such as using it for developing a service or product that uses the IP of the Standard.

There is a variance to the above which are typically called de-facto/semi standards which are defined as specifications which are accepted by its relatively widely spread usage.

So how can one make sure that a standard is a real Standard? One can review it from a “legal” perspective or one could just apply the following logic;

  1. First of all, a real Standard would bear the prefix of the organization who accredited the SDO. for example, the long description of the TIA-942 is ANSI/TIA-942 which means that ANSI is overseeing TIA as an SDO to ensure that whatever they develop is following due process. Just to be clear, ANSI does not validate the content of the standard as this rests with the SDO and their technical committee of SME’s (Subject Matter Experts).
  2. A real Standards (typically) has a numeral indicator e.g. ISO-9001, TIA-942
  3. A real Standard is a document which provides a clear description of all audit criterion

Coming back to the main question and based on the explanation provided I believe it is very clear, and nobody can even argue, that ANSI/TIA-942 is a real Standard. UTI-TST is not a Standard but a guideline. At best, and with a fair amount of imagination, you could consider calling it a de-facto standard but anything beyond that statement clearly is a misrepresentation of the facts and the intent as how WSC and its members would define and recognize an SDO and a Standard.

Article 4 | Uptime vs. TIA-942: What is within the scope?

One of the key differences between UTI:TST and ANSI/TIA-942 is the scope. For the TST topology guideline of UTI the scope is very clear as it only covers the mechanical and electrical infrastructure. This is often seen as inadequate by data centers owners. As one of a data centre consultant once said to me “you could build a data centre in a wooden hut next to the railroad track and nuclear power plant with no fire suppression and the doors wide open and still be a Tier-IV data centre based on UTI:TST. As ridiculous as it might sound, the reality is that nobody could argue with this consultant as the UTI:TST only covers electrical and mechanical, full stop. Although electrical and mechanical systems are very important, it doesn’t make any sense to ignore all other aspects that would contribute to a reliable, secure and safe data centre.

For ANSI/TIA-942 the situation is slightly more complicated. Officially the standard is called “Telecommunications Infrastructure Standard for Data Centers”. There are a number of annexes in the ANSI/TIA-942  which describe additional criterion such as site location, building construction, electrical and mechanical infrastructure, physical security, safety, fire detection and suppression etc. So, one could easily figure out that ANSI/TIA-942 is clearly covering all aspects of a data center. So what is the issue?

There is a theoretical and practical side to this. Let’s start with the theoretical side first. The standard indicates in the introduction that the 8 annexes are not part of the requirements of the standard and as such the annexes start with the term ‘informative’. However, a few sentences later it states “It is intended for use by designers who need a comprehensive understanding of the data center design, including the facility planning, the cabling system, and the network design”. This indicates that the Technical Committee who put the standard together has a clear intent to cover the whole data centre and not just the network infrastructure alone. Furthermore, the standard also states that “Failsafe power, environmental controls and fire suppression, and system redundancy and security are also common requirements to facilities that serve both the private and public domain’. In addition to this, in Annex-F it states “This Standard includes four ratings relating to various levels of resiliency of the data center facility infrastructure”.  It is hard to ignore by the continues reference to the relation between telecommunications and facilities infrastructure that this should be taken as an overall design standard and not just for telecommunications alone.

Then we have the practical side of the matter which is that any data centre which is taking the ANSI/TIA-942 as their refence point does so by referring to Tier/Rating levels. I have never seen any data centre declaring conformity to the ANSI/TIA-942 which ignored all the annexes as by right one could then just pull the approved network cables in the right way and forget about all other aspects such as electrical and mechanical systems etc. The reality is that data centre operator/owners who are using the ANSI/TIA-942 standard as their reference point are using its full content, including the annexes and rating systems.

So, the conclusion is very simple. No matter how much confusion some parties try to throw into the mix, the reality is that data centre designers/operators/owners take the full document as their reference for designing and building a reliable, secure, efficient and safe data centre. Anybody who says that ANSI/TIA-942 is only used for telecommunications  is either ignoring what is happening in the real world or is just oblivious to the facts of how ANSI/TIA-942 is being written and/or used.

Article 5 | Uptime vs. TIA-942: Outcome based or checklist or can it be both?

In this article, in the series of articles about Uptime vs. TIA-942, I will address a statement often used in favour of Uptime vs. TIA-942. Consultants favouring Uptime are typically using the argument that they are not using a checklist but are assessing designs based on desired outcome. The claim is that ANSI/TIA-942 is not flexible and prevents innovation of designs as it is using a checklist i.e. tick in the box approach. So, let’s examine the true facts of these statements.

Checklists based:

First of all, UTI does have a checklist. However, it is an internal checklist which is used by their own engineers to go through designs in a systematic way. This checklist is not shared with the general public, even though it would be helpful for everybody to have it, in order to get a better understanding about the details of the UTI demonstration/test criteria. This goes back to one of my previous articles about what real standards are, i.e. open and transparent.

ANSI/TIA-942 is a combination of descriptions of what needs to be achieved to meet defined rating levels as well as supplemental annexes to provide guidance on how to achieve this. However, make no mistake, purely applying the table of annex-F as a checklist for conformity without considering the rest of the standard will give you an ugly surprise during an audit as the table is a supporting element to the standard, it is not intended to be a complete checklist for all requirements of the standard. This is a classic mistake of inexperienced consultants/auditors offering consulting/audit services and proudly pull out a copy of the table, putting a tick in every box and then declare a site to conform to ANSI/TIA-942.  These consultants/auditors have clearly not understood the standard and/or do not understand how audits should be conducted. Unfortunately in EPI we have seen data centre owners in “tears” when during an audit we found major non-conformities which were overlooked by these kinds of consultants. Be aware of whom you choose for consulting and audit engagements and make sure they apply the ANSI/TIA-942 appropriately.

Outcome based:

This applies to both UTI and ANSI/TIA-942. Don’t forget that ultimately the description of what constitutes to be a Tier:I-II-III-IV is exactly the same as what ANSI/TIA-942 describes as Rated:1-2-3-4.  For example, UTI:TST defines Tier-III as a Concurrently Maintainable (CM) data centre, similar to ANSI/TIA-942 defining a Rated-3 as a Concurrently Maintainable data centre. However, from the previous article you have learned that UTI:TST only covers Electrical and Mechanical (cooling) whereas ANSI/TIA-942 also includes requirements for Telecommunications to meet the CM requirement. A key difference is of course that ANSI/TIA-942 provides much more transparency and guidance on how this could be achieved, by giving clear indications of what one should/could do to achieve this.

Here is an example of how it works in the real world, which is very different than what is portrayed by consultants favouring Uptime. ANSI/TIA-942 states that for a Rated-3 data centre there should be 2 utility feeds which can come from a single substation. What about if you have only one utility feed? You could still meet Rated-3 if you can prove that you meet the overarching statement of being CM. So, if you have generators and you can prove that during planned maintenance you can switch to the generators then you could still be meeting Rated-3 requirements. Of course, there will be a number of other criteria which you will need to address to ensure that the generator is capable of continuous support of the load over extended period of time etc. but in essence you certainly can meet Rated-3 despite not having followed the exact word in the table that says you need two utility feeds. Inexperienced consultants/auditors do not understand this, Certified consultants/auditors will understand this so make sure you put your design work in capable hands. Consultants favouring Uptime often will try to scare the customer with the “you will never be able to comply to ANSI/TIA-942 as the table tells you that you must have XYZ”. It is kind of hilarious to see that the same type of consultants declare that the annexes are not part of the ANSI/TIA-942 standard but yet try to scare a customer of not meeting the items listed in the very same table they say are not part of the standard…

So, coming back to the question of “Outcome based or checklist or can it be both?”  UTI is outcome based and does not provide practical guidance on how to achieve it. ANSI/TIA-942 is also outcome based but provides guidance by means of clear descriptions in various annexes and a supplemental table for you to use. Those with more advanced technical skills can still use the flexibility to implement the design differently as long as they meet the outcome objectives and guidance of the annexes. Therefore, there is absolutely no truth to the statement that ANSI/TIA-942 hinders innovation when designing data centres.

Feel free to share this range of articles to other LinkedIn groups, friends and other social media.

In my next article, I will address the often-heard misconception of “Uptime is easy, TIA is hard”

LikedUnlikeUptime vs. TIA-942: Outcome based or checklist or can it be both?CommentShareShare Uptime vs. TIA-942: Outcome based or checklist or can it be both?


Article 6 | Uptime vs. TIA-942: Uptime certification is easy, ANSI/TIA-942 certification is difficult

Following up on my previous article, we will now have a closer look at a statement which consultants favouring Uptime tend to throw at data centre operator/owners trying to convince them to go for Uptime certification instead of ANSI/TIA-942 certification. One of the famous statements is “Uptime certification is easier to achieve compared to TIA-942”.

When hearing this, the first thought that comes to my mind is what my late father always said, ‘If something is too easy to achieve, it is probably not worth it’. Having said that, I don’t think that Uptime certification is all that easy to comply with. So why is it that such statements are being made?

At the core of those using such statements is of course a commercial interest by trying to scare data centre owners/operators from pursuing the ANSI/TIA-942 certification. So, what are their justifications and are these true or false? The arguments usually brought up in those conversations are;

  1. UTI:TST only reviews electrical and mechanical (cooling) systems whereas ANSI/TIA-942 is too complicated as it covers everything including telecommunications, physical security etc.
  2. ANSI/TIA-942 is prescriptive and has many strict requirements in the Table which are hard to implement
  3. If you fail to meet one of the ANSI/TIA-942 requirement then you cannot get certified

Let’s have a look at each of these statements one by one to decipher the truth;


Argument-1; UTI:TST only reviews electrical and mechanical;

There are two items to be examined here;

  1. The scope of the audit, in this case being electrical and mechanical
  2. The difficulty in meeting the criteria for the defined scope

As for the scope, yes it is true that the smaller the scope is, the less will be assessed and therefore potentially less issues might be discovered. But to me that sounds like saying that you have a safe car because your seatbelts are certified but you didn’t look at the tires, the structural strength of the car and other factors that have an impact on the overall safety. Similar to a data centre, you can review only electrical and mechanical systems like UTI:TST but if you don’t review the network, physical security and other factors then you still have a very large risk at hand from an overall data centre reliability perspective. So, as a business manager running an enterprise or commercial data centre, or as a user of a commercial data centre, would you be happy to know that a certificate is not covering all aspects that potentially poses a business risk to you? A broader scope like what is in the ANSI/TIA-942 will ensure that all potential physical risks are evaluated.

As for the difficulty of meeting the criteria for the defined scope, in this respect UTI is certainly not easier compared to ANSI/TIA-942. In fact, some of the requirements from UTI are considered to be more difficult; an example is the requirement for prime generators whereas ANSI/TIA-942 is allowing standby generators. Another example, UTI is more stringent on ambient conditions as they look at the most extreme condition considering a 20-year history. These two facts alone have have many engineers (and business owners) baffled as it adds greatly to the cost and one often wonders why go to these extremes as ultimately it adds to the cost. ANSI/TIA-942 is in that sense more practical yet allows you to go to these extremes if you wish to do so and are willing to pay the incremental cost. This will give the business an option to choose a well-balance risk vs. investment model.


Argument-2: ANSI/TIA-942 is prescriptive and has many strict requirements in the Table which are hard to implement

This argument is baseless and is aimed at those who do not understand how audits really work. As indicated in one of my previous articles, “Outcome based or checklist or can it be both?” we made it clear that the table is supporting the overarching requirement for each rating level. So, if something is not meeting the exact description of the table it does not mean that you don’t meet the requirement of the standards.  Read the article I wrote about this subject here: “Outcome based or checklist or can it be both?”


Argument-3: If you fail to meet one of the ANSI/TIA-942 requirements then you cannot get certified

This argument pretty much follows the same “logic” of the previous statement. There is NO such truth as not being able to get certified if you miss out on meeting a particular description of the table. Furthermore, in auditing based on ISO, there are Cat-1 and Cat-2 non-conformities. I will explain the difference in a future article but for now it will be sufficient to say that if a site has one (or multiple) Cat-2 non-conformities then that does not automatically mean that a site cannot be certified.


The conclusion is that UTI:TST is being portrayed to be easier based on a narrow scope but that leaves business owners at risk for having an incomplete true assessment of all important factors that make up a reliable data centre infrastructure. If you compare the same scope for UTI:TST vs. ANSI/TIA-942 then both have the same overarching goal such as concurrently maintainability and fault tolerance etc.  whereby in fact UTI can turn out to be more costly due to some requirements which some data centre operator owners consider to be overkill.

In our next article, I will address the usage of the term Tier and Rating. Stay tuned.