How Servers Share an IP Address for High Availability

Cisco's 2017 patent describes a system where multiple servers can share a single virtual IP address to ensure an application stays online even if one server fails.

What it covers

This patent details a system where servers work together to keep an application running. A 'profile' for an application is shared between servers, and this profile includes a virtual IP address (VIP) that clients use to connect. Each server has a 'priority' level. The system automatically picks the server with the highest priority to handle client requests. If the active server fails, another server with a high enough priority takes over. The system constantly checks if servers are still 'neighbors' by sending and receiving status messages. It also tracks how many other servers are sending status messages to each server. If the counts don't match, priorities might be adjusted. For example, if Server A is supposed to get status messages from 3 other servers but only gets them from 2, its priority might decrease.

What it doesn't cover

• —Systems where servers are not at least one router hop away from each other.
• —Protocols that do not use a virtual IP address shared between servers.
• —Systems that do not involve comparing priorities between servers to select an active one.
• —Methods that do not monitor server health via status messages.
• —Applications that do not require high availability.

The clever bit

The innovation lies in how servers actively manage each other's status and priorities, even across multiple network hops. It's not just about one server failing, but about a dynamic system where servers collaborate to maintain a single point of access (the virtual IP) for an application, adapting to changing network conditions and server availability.

Why it matters

This patent is part of the foundation for ensuring critical applications and services remain accessible. It addresses the need for 'high availability' in networking, meaning services should be continuously operational. This is crucial for everything from online banking to cloud services, where downtime can be very costly.

Real-world examples

• 1.High-availability clusters for critical servers
• 2.Network load balancing solutions
• 3.Virtual IP address management in data centers