Reading a datacenter facility record: what the public data actually tells you

Every field in a datacenter facility record belongs to one of two categories: data that can be independently verified against a source that exists regardless of what the operator says, and data that reflects what the operator has chosen to disclose. These two categories are not always clearly labelled, and the presentation of a facility record in a directory or on an operator's own website typically treats both as equally authoritative. The practical difference in how much weight to assign each type of information is significant.

Independently verifiable fields include the autonomous system number and the entity that holds the allocation, the address blocks announced from the facility, the exchange point memberships confirmed for networks physically present at the location, the carrier cross-connect relationships documented in the internet registry and exchange peering databases, and any certifications whose currency and scope can be confirmed against the issuing body's registry. These fields tell you something about the facility that the operator cannot misrepresent in the public record without creating a discrepancy that would be visible to anyone who checks.

Operator-reported fields include physical capacity specifications such as total power available in megawatts, cooling design and redundancy architecture, total floor space, available cage and rack inventory, remote-hands staffing, and SLA terms. These fields can be accurate and are often the most practically useful for a procurement decision, but they require either a trust relationship with the operator or a site visit to verify independently. The distinction matters most when the operator-reported fields are the basis for a significant financial commitment or an insurance policy condition.

The autonomous system number and the entity that holds the allocation are the most fundamental independent signal in a facility record. The autonomous system registration, maintained by the relevant regional internet registry, names the legal entity that applied for and holds the allocation. That entity is the network operator, and confirming that it matches or is clearly traceable to the colocation operator you are evaluating is the first verification step. When the autonomous system holder is a different entity than the one presenting itself as the datacenter operator, the discrepancy may be explained by a subsidiary structure, a rebranding, or a managed-services arrangement, but it warrants a specific question before you proceed.

The announced address blocks tell you the scale of the operator's own network footprint. An operator announcing a large number of prefixes from a facility has deployed substantial network infrastructure there. An operator with no announced prefixes at all is likely a pure real-estate colocation provider rather than a network operator, which is not a negative signal in itself but means the connectivity you access through the facility will come entirely from the carriers present rather than from any owned network infrastructure the operator brings. This distinction is relevant for evaluating the carrier neutrality of the facility in a meaningful sense rather than simply as a marketing label.

The exchange point membership records confirm which networks are physically present in the building as active exchange participants. This is the most reliable carrier diversity signal in a facility record because exchange membership requires physical equipment at the location and active participation in live traffic exchange, both of which can be confirmed through the exchange point's own membership directory. A facility that hosts a large exchange with diverse membership has a demonstrably different connectivity profile from one that does not, and this fact is confirmed by an independent third party rather than asserted by the operator.

The certification fields in a facility record tell you the most when they include the certificate reference number alongside the certification name, because the reference number is what allows you to look up the certificate in the issuing body's registry and confirm its current status, expiry date, and scope. Without the reference number, a certification field tells you only that the operator claims to hold or to have held a certificate of a particular type at some point, which is a much weaker signal than a confirmed current certificate with a defined scope covering the hall or infrastructure you are assessing.

The scope question is particularly important for ISO 27001 and SOC 2 certifications. ISO 27001 covers the management system within a defined scope that is set at the time of certification and may or may not encompass all of the operator's facilities and services. A large operator with a regional certificate that covers its flagship campus may not have extended that certificate to a newer or smaller facility in the record you are reading, and the certificate will not say this explicitly unless you look at the scope statement in the issuing body's registry. SOC 2 Type II reports cover a specific audit period, typically six to twelve months, and the most recent report may be twelve to eighteen months old by the time it is referenced in a facility record. The recency question is worth asking explicitly: when was the last audit, when is the next one scheduled, and is the auditor willing to provide a bridge letter confirming no material changes in controls since the last report?

Uptime Institute Tier certifications are the most standardised of the common facility certifications in terms of what the certificate tells you. A current Tier III or Tier IV Certification of Constructed Facility for a specific building tells you that the physical infrastructure of that building, as built, was assessed by Uptime Institute engineers and found to meet the redundancy and concurrently maintainable standards for that tier. That is a meaningful and specific statement about the physical infrastructure, and it is one of the few certification signals in a facility record that directly predicts availability risk rather than management process maturity.

Physical capacity fields such as total power in megawatts, cooling approach and redundancy level, and total floor space in square feet are almost always sourced from the operator's own disclosure. They are among the most practically useful fields for a procurement buyer and also the least independently verifiable without a site visit or engineering review. The most common errors in these fields are aggregate figures that represent a facility's total capacity rather than available capacity in the specific hall or phase relevant to your evaluation, historical figures that have not been updated to reflect current inventory, and power figures that refer to the utility feed to the building rather than to the power available per rack at the point of use.

The most reliable way to use physical capacity fields in a procurement shortlisting process is to treat them as directional rather than definitive. A facility reporting 50 megawatts of total power is at a different scale from one reporting 5 megawatts, and that difference is likely to survive a verification process even if the exact number does not. The more specific the question you ask of these fields, the more likely you are to need operator documentation to answer it reliably. What is the available power per rack in the specific hall being considered? What is the fully-loaded cooling design temperature, and how has the facility performed against it during peak summer load? These are the questions whose answers distinguish a serious evaluation from one that treats spec-sheet numbers as substitutes for operational evidence.

The most effective way to use a facility record in a procurement process is as a framework for the questions you bring to the operator rather than as a self-contained evaluation. The independently verifiable fields give you a factual baseline that you should confirm matches the operator's own representations before you engage. Any discrepancy between what the public record shows and what the operator says, whether it is a different autonomous system holder, an exchange membership that the operator did not mention, or a certification that appears in the operator's materials but cannot be confirmed against the issuing registry, is worth understanding before you proceed to a site visit or an RFP.

The operator-reported fields give you the agenda for the site visit and the documentation request list for the technical evaluation. For each physical capacity specification in the record, the due diligence question is: what documentation can you provide to support this figure, covering the specific phase or hall we are evaluating rather than the campus aggregate? For each certification, the question is: can you provide the current certificate with reference number, and can you confirm the scope statement covers the infrastructure relevant to our deployment?

Working through this process systematically before and during vendor engagement produces a documented evaluation that is grounded in independently verified facts rather than in marketing representations. That documentation is valuable internally for procurement governance, externally for compliance audit purposes, and practically for reducing the risk that a significant infrastructure commitment is based on a mischaracterisation of what the facility actually provides. viabandwidth builds each facility record in the directory from this methodology, distinguishing what network evidence confirms from what operators report so that the confidence level of each data point is visible rather than implied.