Disaster Recovery Colocation: How to Choose a DR Site

What disaster recovery colocation is (and is not)

A disaster recovery (DR) colocation site is a second, physically separate facility where you place hardware, storage and network gear so critical systems can keep running, or be restored, when the primary environment is unavailable. The primary may be your own server room, another colocation facility or a public cloud region.

DR colocation is different from facility-level redundancy. A single building with 2N power and a strong Tier design still shares one location, one utility region and one set of local risks. Redundancy inside a building is covered in Colocation Power & Redundancy and Tier 3 vs Tier 4 Data Centers. DR colocation addresses the risk that the whole site, or the whole metro, is out.

This guide is for buyers who have decided they need a second site and now need to rent the right space. If you are moving an existing environment rather than adding a recovery site, see the Colocation Migration Guide.

Start with recovery objectives, not square footage

Your recovery time objective (RTO) and recovery point objective (RPO) decide what kind of DR site you need, which in turn decides how much space, power and connectivity to quote.

Write down the RTO and RPO for each application tier before you request quotes. It is common for only a subset of systems to need hot standby while the rest can recover from backup, and that mix lowers the committed footprint at the DR site.

How to choose the DR market

Separation from the primary

The DR site should not share the primary's most likely failure causes. Consider regional weather and natural-hazard exposure, the electric utility and transmission region, major fiber routes, and staff travel. Too close, and a regional event takes out both sites. Too far, and synchronous replication may not be practical and staff access during an event becomes harder.

There is no single correct distance. Match it to your replication method: synchronous replication needs low latency between sites, while asynchronous replication tolerates more distance. Ask your network and storage vendors what latency your replication design supports, then shortlist markets that fit.

Common primary/DR pairings to evaluate

Buyers often evaluate a DR market in a different region from the primary. For example, a primary in Ashburn might be paired with Chicago or Dallas; a West Coast primary might look inland at Phoenix, Denver or Salt Lake City. These are starting points, not recommendations. Validate hazard exposure, carrier routes and capacity for your own design.

Network path diversity

A DR site is only useful if you can reach it when the primary is down. Confirm that the carriers serving the DR building use physically diverse routes from those serving the primary, and that users, offices and cloud environments can reach the DR site independently. Carrier-neutral facilities make this easier because you can select carriers per site; see Carrier-Neutral Data Center.

If you are torn between two candidates, read our guide to choosing between two metros.

Smaller metros are often a good fit for a second site, see secondary market colocation.

What to quote for a DR site

Power and space

Quote the power you need on day one and the power you would need during a full failover. For cold or warm standby, ask for reserved expansion capacity in the same hall or cage rather than paying for full capacity you rarely use. Get the scale-up terms in writing: how much notice, how quickly power can be energized, and at what commercial terms.

Connectivity between sites

Plan the links that carry replication traffic, management access and failover traffic. Specify the number of cross-connects, carrier or wavelength options, and cloud on-ramps if your recovery plan uses cloud resources. Bandwidth needs at a DR site can rise sharply during failover, so ask how burst or upgrades are handled.

Remote hands and access

During an event, your staff may not be able to travel. Remote hands becomes part of your recovery plan: power cycling, cabling, media swaps, receiving replacement hardware. Define what the provider will do, how fast, and at what rates, especially after hours. Remote Hands vs. Smart Hands explains how to scope this.

Contract terms that matter more for DR

• Expansion and failover capacity rights with defined timelines.
• Emergency access procedures for staff and vendors, including after-hours entry.
• Shipping and receiving for rapid hardware replacement.
• SLA measurement and remedies that apply to the DR site, not just the primary.
• Term flexibility, since DR requirements often change as architecture evolves.

DR site readiness checklist

Use this checklist before you sign and again after installation.

• RTO and RPO documented per application tier
• DR model chosen (cold, warm, hot, active-active) per tier
• DR market separated from the primary's main regional risks
• Replication latency validated against the chosen distance
• Diverse carrier routes confirmed between sites and to users
• Day-one and full-failover power quantified
• Expansion or failover capacity reserved in writing
• Remote hands scope, response times and after-hours rates agreed
• Emergency access list and procedure documented
• Compliance evidence for the DR building collected (if regulated)
• Failover test scheduled and results recorded
• Runbook stored somewhere reachable when the primary is down

Common DR colocation mistakes

• Choosing the DR site for convenience only. A site near the office is easy to visit but may share the primary's utility, weather or fiber risks.
• Quoting DR power as if it were permanent production. Reserve failover capacity instead of paying for full production load you rarely use, unless your design is active-active.
• Single-carrier DR connectivity. If one carrier serves both sites over overlapping routes, a single fiber event can isolate both.
• No written scale-up terms. "We can add power when you need it" is not a commitment during a regional event, when other customers may ask for the same.

Test the plan

A DR site that has never been tested is an assumption. Schedule regular failover tests, record the actual recovery time and data loss, and adjust capacity or connectivity when results miss targets. Include the provider: confirm remote hands and access procedures during at least one test so you know they work when needed.

Before you sign, walk the DR building with a site visit checklist so the space matches the proposal.

Questions to ask providers

1. What is the shortest time from notice to energized additional power in this hall, and can you commit to it in writing?
2. Can you reserve adjacent space and power for failover, and on what terms?
3. Which carriers serve this building, and what route diversity can you document compared with our primary site?
4. Which cloud on-ramps are available in the building or campus?
5. What remote hands tasks do you perform during an emergency, and what are response times and after-hours rates?
6. How do emergency access and escorted vendor visits work outside business hours?
7. What natural-hazard and utility-risk information can you share about this site?
8. How are SLAs measured at this building, and what remedies apply?
9. Can we run a failover test that involves your operations team?
10. What current compliance reports cover this specific building?

Next step

The right DR site depends on your primary location, recovery targets and replication design. DataCenter Hunter can shortlist DR markets and request comparable proposals.

• Get matched DR colocation quotes – share your primary market, RTO/RPO and footprint.
• Book a consult – review site pairing and DR contract terms with a specialist.

Frequently Asked Questions