How to Specify a Containerized Data Center for a Remote or Offshore Site

Short Summary

A containerized data center should be specified as an operating system, not as a container filled with IT racks. Define the IT load and growth path, availability objective, power architecture, cooling concept, environmental conditions, network entry, fire strategy, physical security, transport route, installation interfaces, commissioning tests, and service model before asking for a quotation.

The Short Answer

Choose a containerized data center only after the project has defined what digital services must remain available, the load those services create, and the site conditions in which the package will operate. Racks, UPS units, cooling equipment and an enclosure are only parts of the solution. A workable package must also have a credible power and heat-rejection path, network resilience, controlled access, fire detection and response, transport and lifting arrangements, and enough room to maintain critical equipment safely.

The ISO/IEC 22237 series provides a useful framework for discussing data-center facilities and infrastructure. Part 1 addresses common concepts and a classification approach based on availability, security and energy efficiency. It does not replace the project-specific engineering needed for a remote, marine or offshore installation.

1. Start with the Service, Then the IT Load

Ask what the site will lose if the data center becomes unavailable. The answer may be production data, telecommunications, process visibility, safety support functions, local edge computing, video, access control or a mix of services. Separate life-safety and control-system requirements from general business IT; their availability, cyber-security and change-control requirements may differ.

Create a load schedule that identifies installed and expected IT power, rack density, power factor where relevant, diversity, peak processing periods, growth increments and the heat released by IT equipment. Include network equipment, storage, monitoring, lighting, controls and any local support loads. A room rated by floor area or a container selected by length tells the supplier very little about real cooling and electrical demand.

Define the initial configuration and the design horizon. A package that operates efficiently at 30 kW but is expected to reach 120 kW within two years should be designed around the expansion strategy. State whether the growth will be achieved by spare rack positions, modular UPS blocks, additional cooling capacity, an external plant or a second module.

2. State the Availability Objective in Operational Terms

Avoid requesting “Tier-like” or “high availability” without explaining what that means at the site. Describe the tolerable interruption, recovery time, maintenance conditions, utility reliability, fuel logistics and which single failures must be survived.

Define the power path from source to IT load. The scope may include utility or generator incomers, switchgear, transformer, UPS, batteries, distribution boards, rack PDUs, bypass arrangements, earthing, surge protection and monitoring. State the required ride-through time, generator-start sequence, load shedding, maintenance bypass philosophy and the behavior after a failed UPS module, battery string, cooling unit, breaker or controller.

Redundancy language such as N, N+1 or 2N only becomes useful when its boundary is clear. For example, redundant UPS capacity will not maintain service if both paths depend on one upstream breaker, one room-level control panel, one fuel-transfer system or one cable route. Ask the supplier to identify shared elements and maintenance constraints in the single-line diagram and failure-mode review.

3. Engineer Cooling and Heat Rejection for the Actual Site

Every kilowatt consumed by IT becomes heat that has to leave the module. The cooling system must be selected from the site design conditions, not an annual average weather value. Specify outdoor dry-bulb and humidity ranges, solar exposure, altitude, dust, salt, driven rain, wind, corrosive contaminants, water availability, permitted water treatment and any low-temperature or icing conditions.

For air-cooled systems, confirm the location of condensers or dry coolers, airflow paths, pressure drops, discharge recirculation, salt protection, noise, access and crane-removal route. For liquid or chilled-water systems, define the water-quality boundary, freeze protection, leak detection, pumps, pipe materials, isolation, external connections, drainage and who owns the central plant.

The desired IT inlet conditions, permitted temperature and humidity range, alarm thresholds, lead-lag control, failed-unit behavior and emergency ventilation should appear in the control narrative. If the data hall contains high-density racks, specify the intended rack layout, blanking panels, containment approach, cable routing and allowable airflow obstruction. Cooling redundancy cannot compensate for uncontrolled air bypass inside the module.

4. Treat Connectivity as a Physical and Operational Interface

Remote sites need a connectivity design that is as deliberate as the power design. Provide the carrier or satellite arrangement, bandwidth, latency expectations, network demarcation, fibre or copper entry points, route diversity, cable protection, grounding, surge protection, telecom-room interfaces and requirements for out-of-band management.

Two services from the same physical landing point or mast may not deliver meaningful resilience. Map the route from each external provider to the active network device and record common ducts, cabinets, penetrations, power supplies and environmental risks. Protect cable entries against water, rodents, fire and electromagnetic exposure as appropriate to the site.

Cyber security is not created by the container shell. Agree who owns network hardening, remote access, logging, patching, asset inventory, configuration backups and incident response. The module should provide controlled physical interfaces and environmental monitoring, while the wider system needs an operating model.

5. Define the Enclosure and Safety Boundaries

The enclosure should be selected around the installation and operating environment. Confirm structural support, floor loading, access doors, roof openings, cable and pipe penetrations, weather protection, corrosion system, insulation, vapor control, lighting, emergency lighting, drainage, noise and site-specific wind or seismic criteria where applicable.

Fire strategy must be agreed with the authority having jurisdiction, insurer, operator and relevant project standards. State the detection technologies, alarm interfaces, shutdown cause and effect, suppression approach if required, manual release or abort arrangements, emergency power, signage, evacuation, service access and post-event recovery plan. A suppression system inside the container is not a substitute for site-wide emergency coordination.

Specify access control, CCTV, intrusion alarms, key management and the interface to the site security system. Also plan human factors: safe maintenance clearances, lifting points for replacement equipment, emergency exits, safe isolation points, ergonomic rack access and a clear route for failed batteries, UPS modules and network equipment.

6. Plan Transport, Installation and Commissioning Before Fabrication

Determine whether the package must travel by road, sea, rail or air and define the transport configuration. Dimensions, mass, center of gravity, lifting points, forklift restrictions, sea fastening, shock limits, preservation, humidity control and removable external equipment must be known early. A custom data-center module is not automatically a freight container certified for international container transport merely because it has a container shape.

For an ISO series 1 freight-container transport basis, ISO 1496-1 covers specifications and testing for general-purpose containers, and the International Maritime Organization explains that the CSC addresses testing, approval, inspection and maintenance for most freight containers in international transport. Project-specific equipment modules can have a different certification boundary; clarify this rather than assuming a CSC plate approves the installed IT system or the final site installation.

Prepare a site-interface schedule covering foundations or deck support, anchoring, cable routes, network entries, power incomers, external cooling connections, drainage, access platforms, fire and gas signals, earthing, lightning protection and commissioning authority. Factory acceptance testing should prove the agreed package functions before shipment. Site acceptance testing should then confirm the interfaces and operating scenarios that could not be fully reproduced in the factory.

Information the Buyer Should Provide with the RFQ

Site location, application, installation date and operating life
Critical services, availability objective, recovery expectation and expansion plan
Initial and final IT load, rack count, rack density and equipment heat load
Electrical single-line diagram, supply quality, utility or generator arrangement, autonomy and earthing basis
Ambient conditions, corrosion category, dust, salt, wind, solar load, altitude and water availability
Cooling preference, external plant boundary, noise limit and water-treatment constraints
Network providers, entry points, route-resilience requirements and remote-management approach
Fire, gas, security, hazardous-area and emergency-system interfaces where applicable
Transport route, lifting method, installation support, access restrictions and certification basis
Required FAT, SAT, documentation, training, spares, warranty and service-response expectations

Procurement Takeaway

The right containerized data center is one whose service objectives, power, cooling, connectivity, safety and maintenance design remain credible after it reaches the actual site. TLS can configure containerized data center solutions around the project-defined electrical, mechanical and IT scope. Send the service description, load schedule, site conditions, interface drawings and testing requirements for a focused technical review before fabrication.

FAQ

Can a containerized data center be deployed outdoors?

Yes, if the enclosure, cooling and external interfaces are engineered for the local temperature, humidity, contamination, wind, rain, solar and corrosion conditions. Outdoor placement does not remove the need for foundations or supports, drainage, secure access, cable protection and service clearances.

Does N+1 power capacity guarantee service continuity?

No. It only describes a capacity relationship within a defined boundary. Review shared upstream distribution, control systems, cooling, network paths, fuel systems, cable routes and maintenance procedures to understand the real single points of failure.

Is a standard freight-container approval enough for a data-center module?

No. A transport approval or CSC plate applies to the stated container transport scope. It does not automatically approve installed IT equipment, electrical distribution, cooling, fire systems or the final installation.

What is the most important input for cooling design?

The design IT heat load and the site design conditions. Include the load growth profile, rack arrangement, permitted inlet conditions, outdoor temperature and humidity, contamination, available utilities and the location of external heat rejection equipment.