Impact of the Evolution of the Internet on Enterprise Bandwidth Control and Management

How did the bandwidth controllers evolve?The founders of the Internet (a few of whom I had the privilege to meet in person and be collaborate with on common projects) designed the Internet protocol with only a few simple, yet powerful principles in mind. Not knowing the future, these founders utilized robustness and minimalistic design themes to minimize limitations on future applications, while incorporating enough structure to avoid the skeleton of the Internet becoming brittle. Several decades later, with an unlimited list of applications riding on the Internet today, several important questions arise. How do these applications flow through the transport protocols of the Internet? How do they interact with the other applications and the dynamic changing conditions of the Internet? How can these interactions be measured, monitored, and controlled within a complex enterprise IT infrastructure to optimize the end-user experience? How can enterprises successfully implement their bandwidth control strategies in the ever-changing Internet.

These answers are not straightforward, however, keeping the same simplistic design principles, one can come up with bandwidth management guidelines and therefore design bandwidth managers that can keep applications under control and provide the required QoS (Quality of Service) that their end users demand.

One key principle of TCP/IP, the primary transport protocol that carries a majority of the data over the Internet, is to optimize bandwidth utilization and be fair to other applications that also use TCP/IP (or similar TCP friendly protocols). The idea is to let flows fill up the available bandwidth and yet make them share the limited resource with other flows fairly. This works close to perfectly as long as all applications agree to abide with the TCP friendliness rules and back off if they go beyond their fair share of bandwidth usage. In reality, because of what happened over the years with custom tweaking of the transport protocols and the IP protocol, the playing field became skewed in favor of some flows that were not TCP friendly.

As an example, the use of the TOS (Type of Service) bit of the header forces routers to give priority to higher-importance packets. Similarly, other types of protocols, unlike TCP (such as UDP), may avoid flow control altogether and rely on the application layer for that application to achieve flow control and TCP friendliness. In some cases, the favoring of one flow might be desired, but in general, the negligence of those design principles by some application developers created applications that don’t behave nicely with other flows. Compound this with the Telcos trying to take bandwidth management in their own hands (dropping / filtering certain types of traffic – also related to the infamous net neutrality discussion) and we ended up in a network soup that has so many ingredients and so many flows that simply don’t play nice with other flows, if there isn’t a cook (bandwidth controller). This created the need for specialized solutions, namely bandwidth controlling solutions.

In other words, bandwidth managers within a corporate network became a necessity in the presence of applications that use protocols other than TCP, that partially or completely ignore the TCP friendliness principles, as well as traffic that has higher QoS requirements and therefore interacts with other flows in different ways.

A good first step is to understand what types of flows are running on your network so that the flows that don’t belong to the network can be weeded out. Most bandwidth management solutions will have this capability to provide you some level of visibility – after all, you cannot manage what you can’t measure. Once you have the capability to filter the flows into specific buckets of traffic types, your bandwidth manager should have the capability to block, traffic shape, limit, and carve out QoS and bandwidth guarantees within the existing total IP pipe. The modern WAN virtualization and SD-WAN¬†approaches will provide the most effective implementations of corporate WANs and of their bandwidth management.

Being able to block traffic is a fundamental component of firewalls and can be done in various simple or more sophisticated ways. Grooming the available bandwidth, i.e. traffic shaping, is crucial and needs to be done in a dynamic manner. In other words, if there is available bandwidth, no application should be punished unnecessarily. If however, there are other more important flows (described by your configuration within the bandwidth manager), the bandwidth controller should dynamically and adaptively limit the flow that was taking more resources than its fair share. A simple example is an implementation of limiting the maximum rate of a file download as soon as real-time traffic (such as VoIP) shows up in the network. When there is no real-time traffic, however, that file download should not be limited unnecessarily.

Fortunately, these type of adaptive and dynamic bandwidth managers are available today. WAN Virtualization devices, such as Broadband Bonding devices, incorporate these features and provide in-depth control of the bandwidth and more importantly, control of the flows that are competing for the limited bandwidth resources.

Cahit Akin, CEO, Mushroom Networks, Inc.

Mushroom Networks is the provider of SD-WAN (Software Defined WAN) and NFV solutions capable of Broadband Bonding that enables self-healing WAN networks that route around network problems such as latency, jitter and packet loss.

https://www.mushroomnetworks.com

Facebook
Twitter
LinkedIn

© 2004 Р2024 Mushroom Networks Inc. All rights reserved.

Let’s chat. Call us at +1 (858) 452-1031 or fill the form: