Dabbawala-WAN-optimization

How is Eating Lunch in Mumbai Like Virtualizing your WAN?

Dabbawala-WAN-optimizationWhat does eating lunch in Mumbai have to do with virtualizing your WAN?

Plenty, it turns out.

Like Mumbai’s fascinating and highly efficient lunch delivery system, WAN optimization software helps your network adapt to real-world business demands, on the fly.

Instead of going home for lunch or paying for a meal in a cafe, many Mumbai office workers have a cooked meal sent from their home. A collecting dabbawala, or lunch worker, usually on bicycle, picks them up. The dabbas have some sort of distinguishing mark on them, such as a color or symbol. Like protocols in a packet-switched network, these codes define the routing and sorting.

Mumbai is a densely packed city with huge flows of traffic. Lengthy commutes to workplaces are common, with many workers traveling by train. The dabbawala takes the lunches to a designated sorting place. The grouped lunches are put in boxcars, with markings to identify their destination. The markings include the rail station and the building address where the container is to be delivered.

When a palette is full of lunches, succeeding lunches might make the next palette or the next railroad car, or even a different train. Humans use judgment to make the system work.

But a typical WAN has no human interlocutors. And now due to the popular implementation of VDI (Virtual Desktop Integration), which stores a virtual desktop on the server rather than locally, demands on the network have never been higher.

VDI adds to system demand because instructions and data from the virtual machine (server) travel the same pathways as data routed over the WAN. So, in off-shoring applications to the cloud there’s more requirement for server cycles and I/O capacity.

In response, WAN optimization allows enterprises and cloud service providers to support more users with existing bandwidth (i.e., more lunches delivered to more people by the same dabbawala), and provides a far better user experience when operating VDI applications over the network (i.e., the lunches still arrive hot).

Existing protocols that support VDI do provide some levels of optimization, but these compression techniques are not suitable for transferring large files, including audio, video and graphics-intensive files. The size of these files, when traveling over the WAN, cause latency and packet loss that can have a significant negative impact on the user experience.

As VDI deployments increase, that consistent user experience becomes more difficult to deliver. Remote users are often connected through different types of WANs, with varying levels of bandwidth, latency, and quality. The greater the distance the users are from the data center, the harder it is to ensure reliability and performance. Supporting thousands of simultaneous sessions, data centers are presented with a scalability challenge.

WAN optimization within a virtual machine infrastructure enables IT managers to meet their consolidation goals and at the same time reap the benefits of lower application response time and network bandwidth utilization.

About the author
Emily Cohen