Request an intervention

Operational notes Regulation

The control room used to have a class of its own. Since May, that class doesn’t exist

7 min read

Perforated front panel of a rack-mounted server seen at an angle, with other blurred equipment panels in the background, black and white photograph
The air moving through these holes has to stay inside a range written into a standard — since May 2026 under a different number, not different values.

The specification for expanding a telecom exchange’s control room lists, under environmental conditions, ETSI EN 300 019’s Class 3.6: +15/+30 °C, relative humidity from 10% to 75%, copied from a specification written a few years back, when it still matched the current text. Whoever checks it today, 1 September 2026, looks up the edition on the ETSI site and finds no trace of that class: in the edition in force since 13 May, Class 3.6 is gone. It was not renumbered. It disappeared the day the series that classifies environmental conditions for telecommunications equipment went from six classes down to three. The gap matters because a specification is a contractual reference, not a memory exercise: what it cites is what a supplier is bound to meet, or is not.

Six classes become three

The ETSI EN 300 019 series, Environmental Engineering; Environmental conditions and environmental tests for telecommunications equipment, has existed since 1992 as ETS 300 019-1-3. The edition in force since 2014, V2.4.1, split stationary use at weatherprotected locations into six classes, from 3.1 through 3.6 — a structure confirmed, at least, since that 2014 text, five editions and one amendment after the first 1992 publication. On 13 May 2026 a new edition, V3.1.1, was adopted: the classes become three, numbered 3.10, 3.11 and 3.12. This is not a cosmetic renumbering. The annex tracking the change history says so without hedging: “The environmental classes have been revised based on the classification given in IEC 60 721-3-3: 2019. As the new classes are not comparable with the previous classes taken from IEC 60721-3-3: 2002, the new environmental classes have been defined with new numbers.” IEC 60721-3-3 itself is a paid standard: we name it here only because ETSI’s own text does, by title and edition, and we have not read its content.

Before the numbers, it is worth reading the clause that says what a class actually is. The scope of V3.1.1 puts it plainly: “The present document defines classes of environmental conditions and their severities to which telecommunication equipment may be exposed. The severities specified are those which will have a low probability of being exceeded; generally less than 1 % of the operating time in a year.” A class is not an absolute promise. It is a range the room is allowed to exceed, rarely, without that failure meaning the classification was wrong.

Class 3.10: the same air as an office

Of the three new classes, the one that applies to a data hall is 3.10, Temperature-controlled enclosed locations having no humidity control. Table 1 sets the figures: minimum temperature +5 °C, maximum +40 °C under normal conditions; relative humidity from 5% to 85%. Under exceptional conditions — labelled 3.10E, meaning a failure of the climate-control system rather than a separate class — the range widens to −5/+45 °C and 5%-90%. Those are, digit for digit, the same numbers the old Class 3.1 carried, absolute humidity included: 1 to 25 g/m³ either way, a figure specifications rarely restate even though it is what actually separates a controlled room from open air. The data hall has not changed climate; it has only changed number.

What has not changed at all, across the two editions, is the company a data hall keeps. The standard lists where the conditions of this class turn up, and it is the same list it printed twelve years ago: “normal living or working areas, e.g. living rooms, rooms for general use (theatres, restaurants); […] offices; […] shops;” down to “storage rooms for valuable and sensitive products; […] data centres; […] computer halls.” Same class, same temperature and humidity band, for a living room and for a machine hall. The standard even spells out what it means by a data centre: “data centre: all buildings, facilities, offices and rooms which contain enterprise servers, server communication equipment, cooling equipment and power equipment, and provide some form of data service.” Nothing in that definition, or in the class it sits under, singles out a server hall for tighter treatment than a shop.

The class that no longer exists

The old Class 3.6 never covered the whole machine hall: the standard reserved it for something narrower. It read: “dedicated control rooms within telecommunication centres intended to be used for small specialized peripheral equipment such as workstations, test equipment, storage media and printers.” Its range was tighter than a generic data hall’s: +15 °C to +30 °C, relative humidity from 10% to 75%, absolute humidity from 2 to 22 g/m³ — against the 5%-85% and 1-25 g/m³ of Class 3.1/3.10. In the new edition, that class has no successor: whoever still cites 3.6 in a specification signed after 13 May 2026 is citing a number the current edition simply does not contain. V2.4.1 stays downloadable as a historical text, but the withdrawal of any conflicting national standard is due to complete by 28 February 2027. Until then, a specification has to state which edition it is citing, not just which class number.

The proof isn’t a standard yet

A class, on its own, is not a proof. Whoever needs to verify that a piece of equipment holds up to a class turns to Part 2 of the same series — for stationary use at weatherprotected locations, ETSI EN 300 019-2-3. This is where the timing comes apart: Part 1 has been in force since 13 May, but Part 2-3, updated to the new class numbers, is still a draft. Its cover page says so outright: “This draft European Standard (EN) has been produced by ETSI Technical Committee Environmental Engineering (EE), and is now submitted for the combined Public Enquiry and Vote phase of the ETSI EN Approval Procedure (ENAP).” The document’s own history table dates that window from 9 July to 30 September 2026 — as we write this, it is still open, and the final text could still change before it closes. It is the same Public Enquiry and Vote track that Part 1-3 itself went through only months earlier, between February and May 2026, before it came out the other side as V3.1.1.

And even once approved, that standard will not certify a data hall that is already built. It will certify a piece of equipment inside a climatic chamber, before it ever gets installed. The draft says as much in its own setup clause: “The equipment shall be tested in its operational state throughout the test conditions described in the present document unless otherwise stated.” A specification that asks for compliance with Class 3.10 has asked for something the manufacturer can prove on the equipment, in a lab, before delivery. Nothing in the EN 300 019 series says how to verify, a year after switch-on, that the room itself — with its cooling units cycling on and off, exactly as Class 3.10 already assumes they will — is still inside that range.

The point

A written environmental class carries three separate claims, and they are worth pulling apart: the class number, which after 13 May only means something once the edition is stated alongside it; the range that number fixes, unchanged for a data hall and gone without a replacement for the narrower control room; and the proof that the range actually holds, which today exists only for the equipment fresh off the line, not for the room that has been running it for years. It is exactly the kind of distinction we check in a specifications and compliance engagement: not Class 3.10 on its own, but Class 3.10 per ETSI EN 300 019-1-3 V3.1.1 (2026-05) — and, for a hall already in service, a measurement of its own, not the one the manufacturer already ran in a chamber.

The manufacturer’s test report, the edition of the standard named in the specification, and the field measurement of a hall’s own temperature and humidity feed, together with CSIDIA, the group’s other company, into a single map where an AI flags any citation of a class no longer matched by the edition in force. It is the same approach behind the data centres we work with, and behind the availability classes of the EN 50600 series we have already covered. Within the client’s own perimeter: on-premise on stand-alone machines, or dedicated cloud with a data centre in Italy, always under shared management.

Are you writing or updating the environmental clauses of a data hall specification? Talk to an engineer: the site visit is free, and together we check whether the class you wrote down still exists, under what number, in the edition currently in force.

Sources