In this quick post, we will look at the most common types of colocation bandwidth pricing models offered from various colocation providers across the United States.
1) 95th percentile – By far, the 95th percentile is the most predominant colocation bandwidth billing option out there right now. Here is how it works using a typical customer example. A client commits to 10 Mbps monthly usage on an internet feed that is burstable to 100 Mbps. The colocation provider polls their client’s bandwidth every 5 minutes over a given month. Since there are approximately 43,200 minutes in a month (30 days x 1440 minutes/day), the colocation provider will have approximately 8,640 data points to analyze (43,200 minute/5 minute intervals). These data point (in Mbps) are sorted in ascending order from the highest to lowest readings. The top 5% of the list (the highest readings) or 432 data points (8640 x .05) are thrown away and the next highest reading is what you are billed for that month. The 95th percentile billing method is best for companies whose Internet traffic only bursts on occasions (36 hours for a 30 day month) and is not good for companies whose traffic burst often and those who do NOT like to get surprised at the end of the month with an additional bandwidth overage charge at the end of the month.
2) Unmetered Bandwidth Plans – Unmetered bandwidth plans come in several options. Some colocation providers offer 10 Mbps and 100 Mbps unmetered plans. Client’s Internet traffic can vary and they only pay a set fee for either 10 or 100 Mbps with no surprises at the end of the month. Unmetered plans are a great option for some companies but as the old saying goes…”Buyer beware!” If a colocation provider offers 100 Mbps unmetered plan to hundreds or thousands of clients, they better be equipped to handle a heavy load of traffic. Unfortunately some colocation providers “oversubscribe” their network bandwidth and their customers suffer from extreme latency and even downtime. When shopping for an unmetered bandwidth plan, make sure your IT folks perform their “due diligence” in selecting a qualified colocation vendor.
3) Total Transfer or Throughput – Most dedicated server and cloud hosting providers offer some type of throughput bandwidth model which is an actual measurement of traffic in (from you to your server) and out (from your server out to the Internet) of your hosted servers. This is typically measured in Gigabytes (GB) versus Mbps. 330 Gigabytes is approximately equal to 1 Mbps. As an example, a dedicated host will offer a server with 500 Gigabyte of bandwidth allocation. You pay a set fee for that package and you are responsible for any associated bandwidth overages. Once again, dedicated server and cloud buyers need to be cautious and make sure your provider has ample bandwidth set up in their network to handle your expected traffic load.
4) Carrier Neutral Bandwidth – Many colocation providers offer carrier neutral bandwidth options. Just like it sounds, clients can pick whatever carrier they want from the colocation provider’s list of available carriers. Many clients prefer this bandwidth model because they already have contracts in place with specific carriers and they can use their existing IP space from their current bandwidth provider. This model is usually not the best fit for most companies because it limits their traffic to one carrier’s network and bandwidth is capped at whatever speed line is chosen (ie) T1 (1.54 Mbps), T3 (45 Mbps), etc.
The above are the most prevalent colocation bandwidth billing models. Before signing a contract, make sure your IT team takes it time to understand what your bandwidth needs are today and for the future. If you are shopping for a colocation provider, check out our free 10 point checklist on how to find the best colocation provider.