by Scott Tyson, Managing Director, EMEA, Auvik Networks
Our use of the internet—and the sheer volume of data we send and receive over it—has grown in ways that many of us never imagined.
Think back to the glory days of the good ol’ Commodore 64 in the mid 80s, when total global internet traffic reached 15GB per month. By 2014, we saw the average user consuming the same amount. And according to Cisco’s Visual Networking Index, more IP traffic will cross global networks in 2020 than in all prior “internet years” combined—up to the end of 2016—and we can expect more than 28 billion devices and connections to be online.
It’s natural to think that, just as the need for storage capacity increases with additional data, the influx of traffic requires more network capacity. But, that’s not always the case.
If the increase in traffic is for legitimate business reasons (like transferring large business files between regional offices) then yes, you’ll want to invest in additional infrastructure to accommodate your growing business needs and avoid lost productivity or any negative implications to the customer experience.
But, if the culprit is Eric in engineering, downloading games at the office (again), then there are other, less costly ways to address that situation. It becomes a matter of ensuring the right traffic is flowing through the network (perhaps an application ban makes sense) and is given the right priority through a quality of service (QoS) policy, rather than increasing the capacity.
So how do you determine what that traffic is, so you can properly address your capacity needs? That’s where network capacity planning comes into play.
There are many ways to approach it, but the goal is generally the same—make better infrastructure decisions, avoid costly downtime, create a better user experience, and never run out of network bandwidth despite your ever-changing business requirements. You need to understand the maximum capability of current resources and the amount of new resources needed to cater to future requirements.
At a minimum, you’ll want to assess:
- Current potential capacity
- Current traffic utilisation
- Type of traffic flowing through the network
- Near- and long-term business plans to evaluate future needs
Assessing current potential network capacity (your bandwidth baseline)
This is the easiest step because the capacity is fixed, regardless of utilisation. The total capacity is the total available bandwidth (the maximum rate that information can be transferred), so if you have a 3Mbps WAN circuit installed between two sites, this is its potential capacity.
This information is usually readily available if your network documentation is up to date. If it’s not, there are plenty of mapping and documentation tools and services available to help with this.
Assessing current traffic utilisation (quantitative data)
In this step, you want to establish the degree to which devices, applications, protocols, services, etc. are generating traffic today.
Gathering and assessing this data has traditionally involved a lot of manual, error-prone tasks, expensive equipment, and in some cases, a period of downtime to install tools on the network. But some of the newer network management systems take the headaches and guesswork out of mapping, documenting, monitoring, and tracking network performance and utilisation, making it a much faster, less intrusive, and more accurate process.
Assessing the type of traffic flowing through the network (qualitative data)
Arguably the most important step, what you discover here will help you understand what’s causing your network congestion or traffic spikes, and if there’s a legitimate business reason for it or not.
But as more and more traffic—both inside and outside the network—becomes encrypted, this step gets more and more challenging. Until recently, the network data necessary for this assessment was lagging. Based on static, historical reports, this method now lacks the deep, real-time insights needed to understand the types of traffic flowing through today’s networks. Fortunately, this situation is now rapidly changing with the help of machine learning.
For example, Auvik TrafficInsights™enhances NetFlowdata with machine learning to show you who’s on the network, what they’re doing, and where their traffic is going—even when the traffic’s encrypted. Leveraging AI to analyse multiple data sources offers greater accuracy than what you’d get with traditional tools.
Near- and long-term business plans to evaluate future needs
In this final step, you want to consider upcoming projects, programs, and business plans that could tax your existing capacity. Are you planning to increase your employee base significantly? Branching out to more locations? Rolling out new applications? Getting ready for mergers or acquisitions?
Using your current and historical utilisation data, you can estimate, plan, and budget for future requirements.
Making an informed decision
When you have the results of your capacity planning in hand, what comes next is up to you. It may make sense to prioritise a network infrastructure upgrade, or optimise the network with QoS policies, or implement (and enforce!) application bans once and for all.
Whatever you decide, you’ll have all the information you need to do so with confidence.
About Scott Tyson
Based in the UK, Scott Tyson is the EMEA managing director for Auvik Networks. Prior to Auvik, Scott was head of global sales at Inbay, where he drove significant company growth in EMEA, North America, Australia, and New Zealand, and founded the company’s first international office. At one time a professional cricket player, Scott emigrated from his native Australia to the UK in 1998, before starting in channel management in 2002. Since then, he’s held senior roles in both the UK and Australia building out regional and global sales channels for companies such as SpectraLink, AdvaTel, and Mailprotector.