In enterprise networking circles, Border Gateway Protocol (BGP) has an unearned reputation for being an intimidating protocol reserved exclusively for Tier-1 ISPs and hyperscalers. When an organisation decides to deploy two independent fibre carriers for redundancy, mid-market IT teams often hesitate at the thought of configuring external BGP (eBGP) sessions.
In practice, BGP is built on simple, deterministic logic. Stripped of telco jargon, it is fundamentally a reachability language that announces to the internet: "These IP addresses live here; send traffic down this path." You do not need a dedicated CCIE or a rack of high-end enterprise core routers to leverage BGP for multi-carrier resilience.
1. The Myth of the Full Global Routing Table
The most common misconception about enterprise BGP is that your edge routers must download the entire global internet routing table to participate in peering.
As of 2026, the global IPv4 routing table exceeds 1,000,000 routes, and the IPv6 table exceeds 200,000. Downloading full tables from two upstream carriers requires several gigabytes of high-speed RAM and significant CPU overhead just to process routing table recalculations during route churn.
Unless you are an internet service provider or an international CDN, you do not need full BGP tables. Upstream carriers offer tiered BGP peering options:
| Peering Feed Option | Route Count | Hardware Impact | Best Applied For |
|---|---|---|---|
| Default Route Only | 1 route (0.0.0.0/0) | < 100 MB | Active/passive or basic active/active dual-carrier failover without complex edge hardware. |
| Default + Domestic/Direct | ~5,000 to 20,000 routes | ~512 MB | Optimises latency for UK/European destinations while sending rest-of-world traffic to default. |
| Full Internet Routing Table | 1,000,000+ routes | 4 GB+ per BGP peer | ISPs, transit operators, and large content delivery networks making global BGP path decisions. |
For 95% of UK enterprises multi-homing two dedicated leased lines, accepting only a Default Route (0.0.0.0/0) from Carrier A and Carrier B is the most reliable, rock-solid configuration. Your router maintains sub-millisecond failover while running lean on hardware resources.
2. ASNs and BYOIP: Achieving Carrier Independence
When an enterprise orders internet connectivity from a single ISP, that ISP issues IP addresses from their own corporate pool (Provider Aggregatable / PA space). If you ever switch carriers, you must re-IP your firewalls, VPN endpoints, SSL certificates, public DNS records, and external mail servers—a migration project that can take months.
Running BGP unlocks carrier independence through two foundational mechanisms:
- Autonomous System Number (ASN): A globally unique 16-bit or 32-bit identifier registered with regional internet registries (such as RIPE NCC in Europe). Your ASN identifies your organisation as an independent routing entity on the global internet.
- Bring Your Own IP (BYOIP): Owning your own public IP address block (minimum /24 for IPv4, or /48 for IPv6). Because you own the subnet, your edge routers advertise the same IP prefix to multiple independent carriers simultaneously.
If Carrier A suffers a backhaul severance or metro fibre cut, incoming packets immediately route across Carrier B via global BGP convergence. Your public-facing services never change IP addresses, and your external services stay online without DNS propagation delays.
3. Steering Traffic: Inbound vs. Outbound Path Control
BGP traffic flow is asymmetric: how traffic leaves your building is decided by your routers, while how traffic enters your building is decided by external upstream networks. Understanding this distinction is the key to tuning multi-homed links:
- Controlling Outbound Traffic (Local Preference & Weight): You have direct control over which link internal staff and servers use to access the internet. By setting a higher
LOCAL_PREFattribute on Carrier A, all outbound sessions prefer Carrier A as long as the BGP session is established. If Carrier A drops, the router automatically steers outbound flows to Carrier B. - Controlling Inbound Traffic (AS-Path Prepending & Communities): External networks select the path with the shortest AS-Path. To make Carrier B less attractive for inbound traffic, you can artificially lengthen your advertised path by prepending your own ASN multiple times (e.g.
AS65001 AS65001 AS65001). Carriers also support standard BGP Communities that let you signal their upstream routers to adjust their local preference dynamically.
4. BGP vs. SD-WAN: Competitors or Partners?
With the rapid adoption of Software-Defined WAN, many engineers ask: "If we deploy SD-WAN, do we still need BGP?"
BGP and SD-WAN operate at different layers of the networking stack and are designed to work in tandem:
- BGP handles binary reachability: It knows whether a circuit is alive or dead. However, BGP has no native awareness of application-level degradation—it cannot detect if a circuit is experiencing 8% packet loss, 120ms of jitter, or poor Microsoft Teams call quality.
- SD-WAN handles application telemetry: Our managed SD-WAN layer sits on top of physical BGP sessions, measuring jitter and MOS scores in real time to steer traffic across the cleanest available link.
In high-availability enterprise environments, BGP provides the resilient, carrier-agnostic underlay, while SD-WAN acts as the application-aware overlay controller. For more on how these layers integrate, read our guide on SD-WAN vs Leased Lines.
5. Peering Security: RPKI and ROA Validation
Historically, BGP relied on implicit trust, making it vulnerable to accidental route leaks and malicious BGP hijacking. In modern enterprise deployments, routing security is non-negotiable:
- Route Origin Authorisation (ROA): A cryptographically signed digital certificate through RIPE NCC proving that your ASN is legally authorised to originate your IP address blocks.
- RPKI Route Filtering: Upstream transit carriers drop unauthenticated or invalid BGP announcements, protecting your public IP prefix from being redirected by rogue third-party configurations elsewhere on the internet.
How ManagedByUs Coordinates Enterprise BGP Deployments
Navigating carrier-side BGP requirements, peering add-ons, and edge routing configurations can create friction for internal IT teams. Because BGP peering typically involves additional provisioning and licensing charges from wholesale network providers, ManagedByUs works directly with your carriers to evaluate commercial feasibility and avoid unnecessary overheads.
As a carrier-agnostic aggregator, we coordinate the technical delivery with the underlying providers (Openreach, Colt, Virgin Media Business, and Neos Networks), configure the routing policies and route maps on your customer edge hardware (FortiGate, Cisco, or MikroTik), and actively monitor session stability 24/7/365 from our UK Network Operations Centre. Backed by our guaranteed 15-minute response SLA, we ensure your multi-homed edge routes cleanly without pulling your internal team into carrier escalation loops.