A data centre move is the one relocation where the lift is the smallest part of the job. A row of full server cabinets, a SAN array still holding live data, a UPS and its battery bank, the CRAC units that keep the room within tolerance, the PDUs feeding every rack and the structured copper and fibre tying it all together — these come out as a working system and only come back as one if the power-down sequence, the cabling and the re-rack are planned to the minute. Across 22 years and 1,213 recorded moves, Seemleius has handled data centre migration and server room relocation in Dubai, across the wider UAE and into the colocation and Free Zone facilities — from a single comms cabinet to a full hall of racks with its own cooling and power train. Most run against a downtime window: the systems are live on Friday, the floor is clear by Saturday, and the business needs services back before Monday trading. That window is what a data centre migration is really about, and it is engineered at the survey rather than improvised on the night.
Equipment categories handled
Server racks & cabinets
Full and partial 42U cabinets, open frames and enclosed racks. Rolled intact where the route and floor allow, or de-racked unit by unit where access or weight demands it, each device asset-tagged to a rack-elevation drawing.
Blade chassis & servers
Blade enclosures, rack-mount and high-density compute. Blades and modules drawn and logged before transit where the chassis must travel lighter, shock-protected and re-seated to the same slot map at destination.
SAN & storage arrays
Disk shelves, controllers and tape libraries. Arrays powered down to a clean state, drives and cabling indexed, and the array shock-monitored on the road so the data on it lands intact.
UPS & battery banks
Online UPS modules, static switches and VRLA or lithium battery strings. Banks isolated and discharged to a safe transport state, battery weight surveyed for floor loading, and the string reconnected and tested before load returns.
CRAC, CRAH & cooling
Computer-room air conditioning and air-handling units, in-row coolers and condenser lines. Refrigerant recovered by a certified engineer, pipework capped, and the cooling re-established before the first rack draws power.
PDUs, cabling & fibre
Rack and floor PDUs, structured copper, fibre trunks and the network core and switches. Every port photographed and labelled, patch records captured, and fibre re-terminated and light-tested at the new room.
Why a data centre migration is different
Move a single machine and the job is finished when it is levelled and the operator powers it up. A live estate has no single finish line. The order of operations is the whole job: an array shut down out of sequence can corrupt, a UPS dropped without isolating its battery string is a hazard, and a rack re-cabled from memory rather than a record leaves ports crossed and a network that will not converge. Two decisions sit at the front of every plan. The first is live-migration against full-shutdown — whether the business can take a clean outage window, or whether services must be carried in waves so something stays up while the racks travel. The second is rack-roll against de-rack — whether a loaded cabinet can be rolled whole down the route and onto the truck, or whether the weight, the floor or the doorways force a unit-by-unit strip and rebuild.
Cabling is its own discipline, and it is where the hours go. A rack is de-cabled to a labelled patch record, structured copper and fibre trunks are logged port to port, and at destination the room is re-cabled to that record with fibre re-terminated and light-tested rather than guessed at. Power-down sequencing is written before anyone touches a switch: storage and applications quiesced in order, hosts brought down, then the UPS and PDUs isolated, so nothing draws or holds a charge it should not during transit. Where the room runs hot-aisle and cold-aisle containment, that layout is reinstated at the new floor — blanking panels, aisle doors and the rack pitch that makes the cooling work — because a re-rack that ignores airflow will run hot the first warm afternoon. And almost all of it happens against the clock, which is why the cutover is rehearsed at survey rather than hoped for on the night.
Methodology and equipment
The work follows the same backbone as every industrial move we run, set out in full on our machinery moving methodology page, shaped here by the need to migrate a live system rather than re-site a single asset.
Power-down and de-cable. A written sequence quiesces storage and applications in order, shuts hosts down, then isolates UPS, PDUs and the cooling so nothing carries a charge into transit. Each rack is de-cabled to a labelled patch record, copper and fibre logged port to port, and asset tags read off against the rack-elevation drawing so every device has a recorded home.
Asset-tag chain of custody. Before anything leaves the floor, each server, array, switch and cabinet is scanned to an asset register and signed for at every handover — off the rack, onto the truck, off the truck, into the new rack. The register is the document that proves nothing was lost, swapped or left behind, and it is reconciled at destination before sign-off.
Anti-static, shock-monitored transport. Devices are handled with anti-static precautions and crated or strapped against vibration; loaded cabinets travel on air-ride trailers with shock and tilt monitors so the SAN and the servers see no impact the data cannot survive. Rack-roll is used where the route and floor permit; de-rack where weight or access does not.
Re-rack and re-cable. At destination devices return to the same slot map on the elevation drawing, cabinets are levelled and bonded, and the room is re-cabled to the patch record. Fibre is re-terminated and light-tested, copper links verified, and the hot-aisle and cold-aisle containment — blanking, aisle doors, rack pitch — reinstated so airflow returns to design.
Power, cool and commission. The UPS and battery string are reconnected and tested before load returns, PDUs are metered and the cooling re-established before racks draw power. Systems are brought up in the agreed order, the asset register reconciled, and the estate handed back with services restored inside the downtime window.
The site survey, in detail
Nothing is quoted firm until an engineer has walked both the origin room and the destination. For a data centre the survey is longer than for a single asset, because the room is a system. We record the power available at the new floor and the sequence the existing room must be shut down in, the cooling capacity and the refrigerant work a CRAC move needs, and the floor loading at every rest point along the route — a loaded cabinet and a battery bank are heavy, and raised access floors have limits. We measure doorway and corridor clearances for rolling cabinets, lift access between levels, and the goods route a truck can reach. Above all we agree the downtime window: how long services can be down, whether the migration runs as one cutover or in waves, and what must be live before the first rack lands. The output is a written method statement with the power-down and re-rack sequence set out, the cutover crewed and costed, and any colocation or Free Zone approvals flagged. The free survey is the foundation of every quote, and for a UAE site we can usually arrange it within a working day or two of your enquiry.
Two server rooms into one over a weekend cutover
A business consolidating two server rooms into a single larger suite, with the racks live on Friday and services needed back before Monday trading. The coordinator held a live programme against the window: the survey caught a raised-floor loading limit on the destination tiles that needed load-spreading under two battery cabinets, every rack was de-cabled to a labelled patch record before breakdown, and crews ran power-down, transport and re-rack in parallel. The SAN was shut to a clean state and shock-monitored on the road, fibre was re-terminated and light-tested at the new room, and the estate was back online before the Monday open.
Cabinet move into a Dubai colocation facility
A company moving a row of full cabinets out of an in-house room and into a colocation hall, where the facility's gate-pass and access rules had to be met before the racks could enter. The coordinator handled the colocation paperwork alongside the move plan, surveyed the goods route and the destination rack pitch, and ran an asset-tag chain of custody so every device was scanned and signed for at each handover. The cabinets rolled whole where the floor allowed, were re-cabled to the patch record on arrival, and the hot-aisle containment was reinstated before power returned.
UPS and CRAC replacement in a live room
An operator swapping an ageing UPS and a pair of CRAC units in a room that could not go fully dark, so the work ran in waves with the load carried on a temporary feed. The coordinator sequenced the power-down so each UPS module and its battery string was isolated and discharged to a safe transport state before removal, brought a certified engineer in to recover the refrigerant and cap the cooling pipework, and staged the swap so the new units were tested under load before the old ones came out. The room held within tolerance throughout and no rack lost power.
We consolidated two server rooms into one over a single weekend cutover, and the business had to be trading on the Monday. The patch record was what made it work — every cable and fibre was logged port to port before breakdown, so the re-cable at the new room had no guesswork in it. The coordinator ran a live programme against the window and the estate was online before the Monday open.
Moving live storage is not a forklift job — the arrays have to come down clean and arrive with the data intact, or you are looking at a restore from backup. The coordinator had the power-down sequence written before we touched a switch, the SAN travelled shock-monitored on an air-ride trailer, and the asset chain of custody meant every device was scanned and signed for at each handover. Nothing was lost and nothing came up corrupt.
The colocation gate pass for one cabinet came through later than I would have liked and held a truck at the facility entrance for a couple of hours. To be fair, the coordinator had flagged the risk in advance and re-sequenced the build so the crew kept re-racking the cabinets already inside, so we lost no time on the migration itself. The de-cable and re-cable were meticulous and the fibre light-tested clean. I would use them again.
Frequently asked questions — data centre migration
Can the migration run as a rolling move rather than a full shutdown?
Both are possible and the survey decides which fits. A full-shutdown cutover takes a clean outage window — usually a weekend — and is the simplest to plan when the business can stand the downtime. A rolling migration carries services in waves so something stays up while the racks travel, which suits an estate that cannot go fully dark but needs more kit, a temporary feed and a longer programme. We set out the trade-off at survey and crew whichever you choose.
How do you handle cabling and fibre re-termination?
Every rack is de-cabled to a labelled patch record before breakdown, with structured copper and fibre trunks logged port to port and the connections photographed. At destination the room is re-cabled to that record, fibre is re-terminated and light-tested to confirm loss is within spec, and copper links are verified before devices are brought up. Re-cabling from a record rather than from memory is what keeps the network converging first time after the move.
Do you test the UPS and power after the move?
Yes. The UPS modules and battery string are reconnected and tested before any IT load returns, PDUs are metered to confirm phase balance and capacity, and the supply is proven on the new floor before racks draw power. Battery strings are isolated and discharged to a safe transport state for the road leg and brought back into service under test at destination, so the protection is real before the estate goes live.
Do you move cabinets whole, or strip them rack by rack?
It depends on weight, the route and the floor. A loaded cabinet can often be rolled whole down a surveyed route and onto an air-ride trailer, which is faster and keeps the slot map intact. Where the weight exceeds a raised-floor or lift limit, or doorways and corridors are tight, we de-rack device by device against the elevation drawing and rebuild to the same slot map at the new room. The survey makes that call before the day.
How do you reinstate hot-aisle and cold-aisle containment?
The existing containment layout is recorded at survey — rack pitch, blanking panels, aisle doors and the airflow direction — and reinstated at the destination floor so cooling returns to its design. A re-rack that ignores airflow runs hot, so blanking and containment go back as part of the rebuild, and the cooling is re-established and proven before racks draw their full load.
How do you keep track of every device during the move?
Through an asset-tag chain of custody. Each server, array, switch and cabinet is scanned to an asset register before it leaves the floor and signed for at every handover — off the rack, onto the truck, off the truck, into the new rack. The register is reconciled at destination before sign-off, which is the document that proves nothing was lost, swapped or left behind. Where you need a secure-erase or certified data wipe on decommissioned drives, we co-ordinate that with your team and provide the certificate.
What downtime window do you need?
That is agreed at survey and depends on the estate and the cutover you choose. A full-shutdown weekend is common: live on Friday, floor clear by Saturday, services restored before Monday. A rolling migration spreads the work across a longer programme with shorter outages per wave. We size the window honestly against the rack count, how much can move whole, and the power and cooling that must be live before the first rack lands, and we flag anything that could threaten the restart.
Can you move an estate into or out of a Free Zone or colocation facility?
Yes. Colocation halls and Free Zone units — JAFZA, DAFZA, Dubai Internet City and others — have their own gate-pass, access and insurance rules, and we file the documentation before the work date is locked. Where equipment is being imported, we confirm the Dubai Customs HS classification first; IT and electrical equipment sits under HS Chapter 84 and 85, and misclassification is the commonest cause of a clearance hold.
Explore the rest of our machinery moving work
Data centre migration is one strand of our wider machinery moving service, and the engineering detail behind every move is set out on the methodology page. If your equipment is leaving the country, our corridor guides cover exporting machinery and industrial plant to the United Kingdom and to India.
Ready to plan a data centre migration?
Send us the rack list, the destination, and the downtime window. One of our team will come back to you with a survey window and a methodology brief within a working day.
Questions people ask about this service
Anything not covered here? A specialist is happy to talk it through.
Contact usHow do I get a quote for this service?
Use the step-by-step quote form — it takes about a minute — or call us directly. Every quote is free, itemised and with no obligation.
Is my shipment insured?
Transit insurance is available on every service. Your coordinator will explain the cover options when quoting so you can choose what suits the shipment.
How far ahead should I book?
Two to four weeks is comfortable for most jobs, but we regularly handle urgent work. The sooner you reach out, the more flexibility you have on dates.
Do you handle the paperwork and customs?
Yes — where a service involves customs or documentation, we prepare and submit it on your behalf and keep you updated at each stage.
Will I have one point of contact?
Always. A dedicated coordinator owns your job from first quote to completion — no call centre, no being passed around.