If your UC or VoIP system goes down when an ISP link fails, no amount of call quality tuning will save the user experience. True availability starts at the routing layer. In this article, we explain how BGP multihoming helps keep voice and collaboration services reachable—even during outages—by giving enterprises control over how traffic reaches their public IP addresses.
How BGP makes your public IP addresses reachable
Even though BGP is a fundamental part of how the internet works, most of us—whether users or enterprise network engineers—rarely interact with it directly. BGP comes into play when an enterprise has public IPv4 or IPv6 addresses that are used to offer services to the broader internet.
For example, an enterprise can host its UC or VoIP system on-premises in its own data center and assign it a public IP address, making it reachable from anywhere on the internet. That address is used by remote and mobile employees when connecting their UC or VoIP clients to the server, as shown in the diagram below.
In order for those clients to reach the server, its IP address—which is 30.30.30.10 in the diagram—must be advertised in such a way that routing devices across the internet become aware of it. That way, whenever a client from anywhere in the world attempts to connect to the 30.30.30.10 destination, its packets will be routed correctly to that destination.
To achieve this, the enterprise must announce its IP prefixes to the local ISP, and that ISP in turn must advertise them to the rest of the world. The result is that any host attempting to connect to the destination 30.30.30.10 will be able to successfully reach the intended server.
This is the primary function of BGP: to announce and propagate routing information for specific destinations to the rest of the internet.
A note of caution: For an enterprise to set up BGP, it must own the IPv4 or IPv6 addresses that it announces using BGP. These are obtained from the regional internet registry (RIR) responsible for its geographic area. Unauthorized announcements, will not be properly advertised at the very least, and at worst, may be seen as malicious activity leading to escalation and even legal consequences.
Advertising the public IP address of a service in this manner means that the server is accessible from anywhere on the internet at large. As such, appropriate security measures must be implemented to ensure that the server is protected from potential malicious attacks.
How BGP advertises networks across the internet
For a service to be reachable from anywhere on the internet, the enterprise’s network must tell the rest of the internet where that service lives. This is done by establishing a BGP relationship between the enterprise’s edge router and its internet service provider. Through this connection, the enterprise advertises the network where its public-facing services reside.
Rather than advertising individual IP addresses, BGP typically announces a group of addresses together. In our example, the server using the address 30.30.30.10 is part of a larger address block, which is advertised as a single destination. The “/24” shown in the diagram indicates the size of that address block. Once the ISP receives this information, it shares it with other networks across the internet, allowing traffic from anywhere in the world to find its way back to the correct destination.
BGP multihoming: building redundancy with multiple ISPs
Multihoming is a technique used by BGP when an enterprise has a redundant connection to multiple ISPs. For simple internet connectivity, connecting to multiple ISPs is trivial, but what if you are advertising public IP addresses? Through which ISP should you advertise them? The answer to that question will affect the path through which your users will connect to your services, and this is where BGP multihoming design comes in.
BGP multihoming is the practice of connecting a network to two or more upstream ISPs and establishing BGP peerings with them to increase redundancy, reliability, and, in some cases, performance. In such a scenario, the edge router forms a BGP peering with each of the ISP routers, enabling it to manage how BGP routes are advertised and, thus, how incoming traffic to those addresses is managed.
Each connection exchanges routes with the ISPs, allowing traffic to take alternative paths if one link fails. This setup provides greater control over inbound traffic flows through BGP policies. Overall, BGP multihoming helps ensure continuous connectivity and improved resilience for enterprises and their data centers.
The key here is that BGP is not simply a routing protocol that chooses the best path. BGP has many parameters and policies that accompany each advertised route. These policies can control precisely how incoming traffic is distributed between the ISPs as well as under what circumstances alternative routes should be taken. BGP can help load-balance incoming traffic, favor one ISP over another, or route all traffic via one ISP, leaving the other one idle until needed. All of these scenarios can be controlled through BGP peerings and associated routing policies.
Multihoming in cloud and hybrid deployments
So far, we’ve been talking about multihoming in an on-premises scenario, but multihoming is also a key part of cloud-based deployments. The concepts are the same, but the scenarios and infrastructure change.
Just as enterprises must advertise their public IP address space to the internet, cloud service providers also need to announce their own address ranges on an even larger scale. These providers operate massive, globally distributed networks, so consistently advertising their IP prefixes is essential for ensuring reachability and optimal routing from anywhere in the world.
Making sure that IP addresses are advertised reliably and redundantly is critical for anyone operating a cloud-based VoIP or UC service. That’s why using BGP multihoming is more important than ever.
Why it's called multihoming
Now, you may be wondering: Why is the name of the feature “multihoming"? It’s called that because a network like the one described above is given multiple “homes” on the internet. Historically, a public IP network connected to one ISP was considered to have a single “home.” When it connects to two or more ISPs and is advertised from both, it becomes multihomed, meaning it resides in more than one place on the internet topology.
So the term simply means one network with multiple homes (connections) for greater availability and control.
Conclusion
Ensuring reliable IP reachability is fundamental for any UC or VoIP deployment, and BGP multihoming provides the resilience needed to maintain that reliability. A multihomed architecture lets providers significantly reduce downtime and deliver a consistently high-quality communication experience.
You may also like:
Boost UCaaS quality with smarter SD-WAN strategies
How to prepare your network for the demands of AI
Unlock scalable VoIP and UC deployments with VXLAN technology
Comments