The Ultimate Guide to Bandwidth Capacity Planning

The Ultimate Guide to Bandwidth Capacity Planning

Bandwidth Capacity Planning

There is no need for a lengthy explanation for why internet connectivity is, without a doubt, the most critical vein (and artery) for the modern enterprise. Every means of production, productivity, communication, collaboration and data gathering in the modern enterprise requires internet connectivity to function. Without it, we are dead in the water.

Bandwidth Capacity Planning should be a central part of any business strategy, no matter the size or vertical of your company. This isn’t a guide to convince you to buy more than you need; this is a guide to help you buy and plan efficiently without wasting money, but at the same time, avoiding losses in productivity. As the age old saying goes, ‘a failure to plan is a plan for failure.’

Keeping Up with the Bandwidth Joneses

Many IT leaders should be familiar with the famous ‘Moore’s Law’. In 1965 Gordon Moore, the CEO of Intel, predicted doubling of components per integrated circuit, combined with manufacturing efficiency gains, and the downward cost of production of said integrated circuits and therefor overall computing power would double roughly every 12-18 months without increasing cost. This principal has help true to present day, and is the reason why we buy new computers, servers, and smart phones for around the same price every few years that offer massive gains in performance. Good stuff.

Over the last three and a half decades, there has been a lesser known, but equally important ‘Nielsen’s Law’ of internet bandwidth growth, which operates on similar observations. Internet utilization has increased by 50% every year, while we see downward pressure on price per bit per second due to market competition and enhancements in network capacity. This is why we can buy a 1Gbps circuit today in 2019 at the same price one could by a 1.5Mbps T1 circuit 20 years ago. Also, good stuff.

Most businesses sign multiyear terms for connectivity, and as such, I urge IT decision makers to think beyond their perceived needs of today, and focus on purchasing connectivity that addresses their forecasted needs of the next two years at a minimum.

A Parallel Example: Page Load Speeds Vs Ecommerce Revenue

There’s no simple formula for the cost of lost productivity from slow connectivity, but rest assured it hides itself in the cost of payroll every month. In order to demonstrate the importance of bandwidth as it pertains to end user productivity, I’m going to need you to follow me down a fairly lengthy parallel example of how page load speeds can easily predict e-commerce revenues. Before you say “well Ben, we aren’t an e-commerce platform, this doesn’t pertain to me”, understand that this is a very tangible parallel example to show that bandwidth lag frustration is VERY REAL, and is probably affecting your users’ productivity and sanity on a daily basis.

There is a wealth of evidence tying page load speed to revenue and other related performance metrics (conversion rate, bounce rate, website traffic, etc.). It has revealed that visitor behavior is EXTREMELY sensitive to fluctuations in page load speed.

Again, these aren’t random samples of users, these are behaviors exuded by the employees of your company and every other company out there.

53% of mobile users said they would abandon a website if the page took four seconds or more to load.

47% of web users expected a webpage to download in no longer than 2 seconds.

79% of respondents admitted they would be less likely to use a company again if they experienced slow page load speeds.

The average US household in 2018 has a download speed of 96.25Mbps and upload speed of 32.88Mbps shared by an average of 2.6 people. Are these people coming into your workplace having to work off far slower speeds because they are shared with hundreds of other employees?

E-commerce Case studies include:

Shopzilla. The e-commerce provider experienced a 7-12% improvement in conversions from their faster loading web pages.

Google. A 100ms increase in page load time saw search volume drop by 0.2%. As load time increased to 400%, search volume fell by 0.6%. Fractions of seconds matter!

AutoAnything. The online truck parts dealer saw sales go up by 13% and conversions by 9% after they halved their page load time.

If e-commerce sites can directly correlate page load speeds to revenue, then it isn’t a stretch to conclude that slow bandwidth will absolutely correlate to lost productivity and end user frustration. If the internet experience you are providing your users doesn’t match their expectations, they aren’t going to be happy employees. No Hawaiian shirt Fridays or Pizza parties can offset the frustrations of slow connectivity in the workplace. These dollars that are lost in productivity show up in payroll costs, which, for most businesses, are their largest cost of overhead.

Loss of Revenue due to Downtime: An Argument for Quality and Redundancy

A typical business grade internet connection is backed with an uptime guarantee of 99.9-99.99%, but only offer up small credits when uptime guarantees aren’t met. In reality, businesses should plan for multiple day failures on any business grade connection. Additionally, there are a multitude of non-outage type issues that can affect connectivity such as carrier network congestion, interface flapping, packet loss and jitter.

The true cost of an outage or deterioration of connectivity can be calculated by the following formula:

Total losses = Loss of revenue + Loss of productivity + Recovery costs + Intangible costs 

Here is how those different elements can be calculated:

Loss of revenue

This can be calculated per hour or per minute as preferred. First, the company’s total weekly revenue is divided by the number of hours or minutes in a week.

Next, a value known as the impact percentage has to be calculated (or estimated). This is the amount of revenue that is dependent on the network. For e-commerce businesses this is easy since 100% of their sales are network dependent. A bricks and mortar store may make 60% of their revenue through online sales so would use this figure instead for the following formula:

Loss of revenue = Hourly revenue x hours of downtime x impact percentage

Loss of productivity

This calculation is similar to loss of revenue but is focused on the salaried hours lost to downtime.

Not every employee will use the network to the same degree so you will need to divide your employees into groups based on network utilization. For example, the janitor may be in the 0% utilization group while the DevOps team would be in the 100% utilization group.

The formula would be:

Loss of productivity = Average hourly salary x hours of downtime x utilization percentage x number of employees in the utilization group

You would then repeat this calculation for each of the groups.

As it takes on average 23 minutes to refocus on a task after disruption, hours of downtime are best rounded up.

Recovery costs

These include repair services (labor and parts), data recovery and other direct data loss costs (compensation, fines, etc.)

Intangible costs

These can only ever be estimated but include any short or long-term damage to the brand’s reputation, reduced customer loyalty and lost opportunities.

How much do these factors add up to on average?

The cost of business downtime for small and medium-sized businesses, according to Cisco, ranges from $8,500 to $74,000 per hour while Gartner puts the overall average at around $300,000 per hour.

The effect of sub-optimal network speeds are similar to that of downtime due to the percentage of visitors who will leave the site or application if it fails to load in an acceptable time.

Bandwidth Capacity Planning and Business Strategy

It should now be clear to see why Bandwidth Capacity Planning has to be treated as an essential part of the overall business development strategy and the necessary executive buy-in and investment secured.

Bandwidth Capacity Planning is proactive, anticipating future business needs and ensuring the network can cope when an increase in demand hits. Extra demand on bandwidth can come from both improvement in the type of technology used (e.g. a new VoIP communication system or media streaming service) and an increase in the number of end users (e.g. following expansion into a new market or a merger).

Redundancy should also be a major consideration, and hopefully by this point in your reading, a no brainer! When compared to the cost of inevitable downtime, redundant connectivity pays for itself many times over! When architecting redundant connectivity, best practice is to use an alternate provider, utilizing an alternate last mile connection that is geo-diverse from the primary connection.

Bandwidth Capacity Planning will also require an audit of current technology. Network flow monitoring tools can be used to spot trends in traffic, identify bandwidth-heavy applications and display the effects of traffic management policies. We at Shamrock Consulting really like LogicMonitor, Nectus5 and SolarWinds.

The Value of Independent Consultation

Due to the many different ways in which a network can be designed to prioritize bandwidth, the help of a vendor-neutral, third party partner is critical. Getting an external consultant onboard for Bandwidth Capacity Planning will ensure you have an overview of all available options.

We provide a free service where we will audit all of your existing services, architect your primary and secondary (or active-active) network, show you the most competitive pricing from all relevant providers, and project manage the implementation of your new network!

Working with Shamrock Consulting Group means that you will have access to the best possible prices which may mean you can afford faster connectivity than you thought. Our slogan is ‘Any Product. Any Provider. Guaranteed best price.’

CLAIM YOUR FREE CONSULTATION
Ben Ferguson

Ben Ferguson

Ben Ferguson is the Senior Network Architect and Vice President of Shamrock Consulting Group, the leader in technical procurement for telecommunications, data communications, data center and cloud services. Since his departure from Biochemical research in 2004, he has built core competencies around enterprise wide area network architecture, high density data center deployments, public and private cloud deployments, and Voice over IP telephony. Ben has designed hundreds of wide area networks for some of the largest companies in the world. When he takes the occasional break from designing networks, he enjoys surfing, golf, working out, trying new restaurants and spending time with his wife Linsey and his dog, Hamilton.