How to Build a Resilient Multi-Site Connectivity Strategy

A multi-site network has to do more than connect offices. It must support cloud applications, voice, remote access, data transfers and day-to-day operations across locations with different availability and risk profiles.

Resilience comes from understanding dependencies and designing around business impact. Buying two circuits is not enough if both services share the same physical route, exchange or power supply.

Start with business requirements

List each site, its users, applications and operating hours. Identify which processes stop during an outage and estimate the effect after one hour, four hours and a full working day. This establishes which locations need dedicated connectivity, rapid failover or a simpler backup.

Match the access service to each site

Different locations can use different products within one overall design:

• Business full fibre or SoGEA can provide cost-effective access for ordinary branch workloads, subject to local availability.
• Leased lines provide dedicated, symmetrical bandwidth and business-grade service commitments for critical sites.
• Layer 2 connectivity can link sites or data centres where private transport is required.
• IP transit and remote peering support networks with their own routing and interconnection requirements.

Exascale leased-line speeds range from 200Mbps to 10Gbps, subject to location, infrastructure and site requirements.

Design genuine path diversity

Two services are only resilient when a single failure is unlikely to affect both. Ask whether circuits use separate carriers, ducts, building entry points, exchanges, data centres and core paths. Document the answer rather than assuming that different product names mean physical diversity.

Plan automatic failover

Define how traffic moves when the primary service fails. The design may use firewall or router health checks, dynamic routing, software-defined WAN policies or provider-managed failover. Test what happens to voice calls, VPN sessions and cloud applications during the transition.

Protect critical hubs

Head offices and data centres often aggregate traffic for other sites, making them potential points of failure. Consider whether branches can reach cloud services directly when a hub is unavailable and whether central security controls have a resilient counterpart.

Colocation in facilities such as Exascale’s Wolverhampton and Telford locations or Telehouse North and West in London can form part of a wider network design, depending on access and interconnection needs.

Use measurable service levels

Compare availability commitments, restoration targets, support hours and escalation processes. Exascale leased lines include a 100% Service Level Agreement with service credits where contractual availability is not achieved, alongside 24/7 telephone and email support from Telford and Wolverhampton.

Monitor the complete service

Router availability alone does not prove that users can reach an application. Monitor circuit state, latency, packet loss, DNS and key services from more than one location. Keep contact details, circuit references and escalation paths in an accessible runbook.

Test failure scenarios

Schedule controlled tests for primary-circuit loss, router failure, power loss and loss of a central site. Record failover time, application behaviour and manual steps. A backup that has never been tested should not be treated as proven resilience.

Review capacity and growth

Track peak utilisation and expected demand from new staff, sites and cloud workloads. Symmetrical bandwidth can be important for video, backups and large transfers. Build an upgrade path before links become consistently constrained.

A practical multi-site checklist

• Classify every site by business impact.
• Map physical and logical dependencies.
• Select suitable primary and backup access services.
• Confirm path, carrier and power diversity.
• Document failover policy and routing.
• Monitor user-facing services.
• Test outages regularly.
• Review capacity after major business changes.

Frequently asked questions

Does every branch need a leased line?

No. Critical hubs may require dedicated service, while smaller sites can use business broadband with an appropriate backup.

Are two internet providers always resilient?

No. Providers may share last-mile infrastructure or building entry routes. Physical diversity should be verified.

How often should failover be tested?

Testing should be regular and aligned with business risk, network changes and operational processes.

Build resilience around the organisation

A resilient design combines the right access services, verified diversity, tested failover and clear operational ownership. Explore business leased lines and other Exascale connectivity solutions for a site-specific design.