Category Archives: Colocation

95th Percentile Methodology: What It Is, How It Works

A term pertaining to colocation bandwidth that you may have heard is the 95th Percentile Monitoring – also known as the 95% Method. So what does the term mean? To illustrate, we will look at a monthly graph of an Internet uplink port:


As you can see, this graph, has the data on the vertical axis, and the pass of time on the horizontal axis. About five weeks’ worth of data is shown here. Please notice that the sample rate has been consolidated to two-hour intervals in this graph – using smaller intervals does not make a statistically significant difference when graphing data over a larger time span.

So, how does a colocation service provider like us compute bandwidth usage for the month using this data? You might assume that it would be fair to add the Average In and Average Out values (Internet traffic travels in both directions) – a total of 36.3 Mbps in this case – and declare that to be the official usage for the month shown. In fact, many other data centers will do just that. However, there is another way: the 95th Percentile Methodology. Here is how it works:

  • A log of 30-days-worth of traffic samples at default settings (typically every 5 minutes) are accumulated;
  • The log is sorted in descending order, which places the highest traffic peaks at the top;
  • The top five percent of the log is discarded;
  • The value of the largest remaining peak becomes the effective bandwidth usage value for the month.

The advantage of the 95th Percentile billing method is that, over the course of a 30-day ‘month’ you can have up to 36 hours of peak traffic and that number is calculated only for the direction of greater flow, making the other direction free (as opposed to the average in + the average out method). In the example of the monthly graph above, the actual 95th Percentile traffic number for that data is 25.7 Mbps – some 10.6 Mbps lower than using the ‘sum of the averages’ method.

As you can see, the 95th Percentile Monitoring can be quite advantageous when acquiring bandwidth.

However, there is one exception to this. That would be in situations where there is normally very little bandwidth activity, but periodically large bursts of high traffic activity that occurs. The 95th Percentile could potentially work against you in the event that the cumulative bursts added up to more than 36 hours in the billing period.

In practice this situation is uncommon; the best way to address this is to try to know (or predict) your usage profile and order service accordingly.

Colocation? Or Keep Your Physical Machines Internal?


Of course many businesses are moving to cloud hosting. However, there is another way that businesses are still using third parties for physical systems: colocation.

  • Considerations for Data Center Cost-Cutting
  • Should You Use Colocation?
  • Controlling Your IT Budget Externally

When you put together a spreadsheet with your data center budget, you really want to be skeptical of every component. You can often nix certain aspects to trim costs, and colocation will often make sense.

Considerations for Data Center Cost-Cutting

Cooling is a huge expense in data centers, and it is easier to control than ever before. You can probably raise the temperature a bit. You can also benefit from a free air cooling or adiabatic system long-term. You will have to pay upfront, but your power bill will drop, and often upkeep will be reduced too.

Is your data center getting bigger? Data center growth is questionable as cloud computing continues to be more heavily adopted as a primary system. The idea of a single app on a singer physical server is no longer reasonable. The old model was incredibly inefficient, with hardware grossly underutilized at 10% server and 30% storage capacity in many cases.

Cloud technology works even when you stay physical, explains Clive Longbottom of TechTarget. “Moving to a flexible, elastic multi-workload private cloud platform should drive server utilization rates up to a minimum of 60%,” he says.

Basically, there is no reason to physically expand when you can make much better use of your current machines through virtualization.

Your data center doesn’t need to be bigger. Does it need to exist?

Should You Use Colocation?

A data center is incredibly costly. First, it’s a piece of real estate. It was not just expensive to build but is expensive to maintain since computing is so fundamental to business and requires built-in redundancies. However, if a company is facing hindrances in setting up a data center due to the lack of knowledge regarding commercial property development, they can look for professional aid. For instance, real estate construction and development experts, such as Lincoln Frost, could be hired by a business to get assistance with property selection, construction, and extracting a higher yield. This can be a great be great way to design a data center that can also be a profitable commercial asset to your business.

As for the functioning of a data center, people with technical backgrounds must be trained to manage the infrastructure. Plus, it isn’t an agile choice. Unlike the cloud, a data center cannot be immediately dissolved when transitioning to a more virtualized system. First, a proper plot needs to be acquired (likely with the help of London Estate Agents). Further, a whole bunch of engineers need to be deployed for the smooth transitioning of each component, be it in terms of machinery or software.

Consider colocation, argues Longbottom. “You retain full ownership and responsibility for the IT equipment, but the facility owner takes responsibility for all that grunt work,” he says.”[T]hey look after power to your equipment, ensure that there is enough cooling and maintain connectivity to the outside world.”

It’s also easier to handle physical security simply because there are so many customers and the costs of monitoring can be distributed throughout all parties.

While your own data center has set parameters, you should be able to easily adjust your space needs in a colocation environment. Now, that’s partially dependent on your provider. You want a hosting company that will make sure they are prepared for your potential growth.

It’s not necessary for your company to have expertise in data center management. In fact, you don’t even have to be focusing on skills in the maintenance of infrastructural hardware. When you take a “cloud and colocation” perspective, you want to focus more in terms of best completing tasks, integrating various elements, and delivering the strongest possible user experience. That’s the sea change of the third platform.

Controlling Your IT Budget Externally

It’s only the first step to decide if you want to try colocation. You want to be gearing yourself toward a diversified, loosely coupled infrastructure that takes advantage of the ecosystems created by third parties.

Public cloud is becoming a more central part of new development, notes Longbottom.

*Related* The key concern with public cloud is that you use a provider delivering the technology “as it was meant to be.” At Superb, we use distributed rather than centralized storage (the latter wrongly used by many cloud providers), so that even multiple node failures don’t affect performance or data. We also leverage InfiniBand for networking instead of 10 Gigabit Ethernet (10 GigE), because the former has dozens of times lower latency than the theoretical minimum of the latter. Get a Superb cloud VM.

You also want to explore software-as-a-service options. In those cases, you are allowing the outside tech company to handle the entire platform.

Along with using SaaS to handle some basic day-to-day tasks, you can manage the security of your virtual and physical systems through cloud security-as-a-service companies.

You can also used managed platforms for development and testing of your new applications.

Are you wanting to work with your own programs rather than do SaaS? It may still make sense to get a cloud server to run it.

Your budget for 2016 should be geared toward eventual removal of the data center entirely in the years to come. You want to be spending in ways that will allow an easier transition completely away from your own facility and toward environments that embrace agility as a core value, Longbottom remarks. “Invest in IT tools that enable greater workload portability,” he says, “in systems that enable ‘what if?’ scenario planning around end-to-end performance management in real time.”

Best Way to Protect Data: Live Inside Your Hybrid Cloud

We all feel ourselves getting sucked into the soul of technology at times, the glow of the screen summoning us into the light, accompanied by a siren song of zeros and ones – that, like the ominous notes of Igor Stravinsky in the theme song of Jaws, seem to pull us into a windowless digital abyss in which, rather than our avatars representing us, we represent our avatars. Sure, getting 100% of your life energy vacuumed into the Matrix may seem unsettling, but as it turns out, a recent study suggests that you want to be as close as possible to some parts of your hybrid cloud.

Debutante Ball for the Hybrid Cloud

Apparently the hybrid cloud deserves an introduction, as indicated by the title of an August 18 Forbes piece, “Get Ready for Hybrid Cloud” (as if it hasn’t been a standard IT setup for years). Well, maybe an introduction is in order. The concept is certainly becoming much more prominent this year, with Gartner predicting that almost half of large enterprises will have active hybrid clouds by 2017.

OK… what is it? Richardson Seroter, writing in Forbes, describes a hybrid solution as a composite of two different choices that allow us to experience the positive elements of both (such as hybrid cars, which combine the sustainability of electric power with the convenience of gasoline refueling). In the case of a hybrid cloud, a business is – in most cases – combining the benefits of a private cloud (proximity, control, security) with the benefits of a public cloud (cost-effectiveness, speed, flexibility). A community cloud can also be part of the mix, according to the US government’s National Institute of Standards and Technology (NIST).

Not everyone agrees on the governmental description above, though, reports Seroter. James Staten of Forrester defines a hybrid cloud as a cloud service that is “connected to any other corporate resource.” Staten’s picture is incredibly broad: it seems to suggest that I could claim I am running a hybrid cloud because iCloud is installed on my laptop. However, it is fair to say that even industry experts are still figuring out exactly what a hybrid cloud is. If anything, it involves slicing and dicing the various cloud technologies to meet custom needs of companies, as opposed to full commitment to one data model, infrastructural architecture, and/or location.

Why Did I Bold That Last Part?

Location, location, location: it’s a good reason not to buy swampland from a telemarketer. Heck, if you’re going to live in Florida, why not choose a neighborhood that isn’t overrun by flesh-eating reptiles? You know what, though? Location isn’t just about avoiding alligators. It can also be critical for data, according to a September 2 report by Wikibon CTO David Floyer outlined in Silicon Angle.

Floyer recommends that the location of an application should be determined by the location of data in order for a hybrid cloud to perform as well as it possibly can. Studies by Wikibon have revealed that, despite the chatter that hybrid cloud is more expensive due to its sophistication, the mix-and-match model offers the “highest value” for large enterprises and some midsize organizations. It can be more cost-effective to store huge applications on-site, says Floyer, but the key concern when creating your hybrid cloud is that applications reside physically close to the data they need to access. Location, location, location.

To make his point, Floyer notes that a common method for constructing hybrid clouds has been to maintain mission-critical data on-premise and host applications that sometimes use that data in the cloud. Floyer thinks that approach is unwise: you don’t want applications to have to travel a long way to get the data but to have it directly available – especially for cases in which there is a delay in the transfer of updated data between various locations.

So Where Do You Put Everything?

Rather than letting the nature of apps determine their location, look to the data to determine location, says Floyer. What this means is different for different scenarios. Software as a Service (SaaS) makes sense in some situations, as does a cloud hosted architecture through a provider such as Superb Internet. Floyer also emphasizes that some organizations can benefit from moving mission-critical applications and data to a colocation center, with hyperscale capabilities and immediate access to data streaming through the public Internet.

A major strength of this approach that will not be immediately evident to all readers is the effect on enterprise employees: by carefully constructing a hybrid cloud so that automation is functioning at its highest capacity, you allow those in your IT department to liberate themselves from mundane repetition and turn toward innovation: creating an app that performs all their work for them, so that they can watch reruns of MASH all day, because they love Alan Alda.

Here’s why you specifically want to work with us at Superb to design your hybrid cloud: we take the confusion out of the cloud, and we make our customers happy. Check out these quick ratings.

By Kent Roberts

Image Credit: Squarespace

How to Save Money When Using Colocation Services

The weather around here has been particularly cloudy lately, hasn’t it? Well, there are a lot of great and interesting things to be said about the cloud, and we know that many of you are looking to hear about it from every angle possible, so that’s what we’ve kind of been focusing on lately. But enough is enough already, right? Well, for now it is. We’ll wrap back around to the cloud later, but right now let’s change things up – let’s talk colocation.

If you’re a web-based business that already owns its own servers then you’re in great shape. But if you’ve dropped a nice little chunk of cash on high-tech servers to store all of your data on, you probably don’t want to shove it into some dusty old room somewhere. Scratch that. You definitely don’t want to do that. Trust us. There’s also a good chance that you don’t want to spend ludicrous amounts of money constructing a brand new state-of-the-art server facility. You do, however, want to store your servers in such a place. The solution? Choose colocation.

Colo gets your very own servers – or ones you decided to lease – stored in a world-class data center (DC) with oodles and oodles of space and high-tech, precise temperature controls (managed and the appliances managed by the right AC Repair professionals on the job) and security measures. It’s the perfect solution for web-based companies that want their data protected while keeping it stored on their own servers that are stored in the absolute best facilities around.

Oftentimes, businesses go the colo route in order to tackle precise concerns or limitations of their own data centers. Insufficient power supplies, underpowered or outdated HVAC systems and limited physical space are the biggest reasons. Other firms are on the lookout for lower workload latency, which can be achieved by putting applications and data for them closer to their users in the physical world. Finally, colocation can be used as part of disaster recovery or business continuity plans.

Whatever the reasons a business considers colocation, though, they’re going to want to do the same thing they do whenever they make any other business decision: save money. Some colo providers require long-term contracts be signed, which can be costly. If you’re considering going this route, then your best bet is to review the information below to glean what the biggest costs are and how you can negotiate a deal with a provider at a price point that works for you.

Continue reading How to Save Money When Using Colocation Services

Hallmarks of a Colocation Center Provider You Can Trust: Part III

If you’ve been keeping up with our blog lately then you caught the first two parts of our look at how to choose a great colocation hosting provider. If you haven’t, not to worry – this is the internet, so the first two parts of this series haven’t gone anywhere. Read up on Part I and Part 2 for an exhaustive look at why you need to consider pricing, service levels and a host (admittedly awful pun intended) of other provider services and capabilities before choosing a data center that works best for your business and its needs and goals.

Once you’ve read up on those key points and checked with potential colo hosting providers as to how they stack up in those areas, you might think you’re ready to make your choice. That’s great if you are, but there’s one more thing to consider: what is their actual facility like?

We hear what you’re saying: “But Superb, didn’t you already tell me to read up and ask around about facility conditions and capabilities?” Yes, indeed we did, and we’re happy to know you’ve been paying attention to our suggestions and taking them seriously. As you should be, since they were designed to help you select the best provider possible. So if you followed through on that one, then you’re probably in a good position with a shortlisted group of hosting names to consider.

Here’s a question for you, though: You wouldn’t buy a car or a house without first looking at it in person, would you? Of course you wouldn’t, and neither should you buy colo center space and services before you take a physical tour of the facility. Obviously you don’t have the time (or likely the desire) to visit every single colo location that sounds decent, but you might want to get out and visit the few in person you’re seriously considering once you’ve narrowed your list of potential vendors down.

When you get there, do you know what you should look for? Specifically? We do. Here are some of the top things you should concern yourself with when you make a physical visit to a data center.

Continue reading Hallmarks of a Colocation Center Provider You Can Trust: Part III

Hallmarks of a Colocation Center Provider You Can Trust: Part II

In case you missed it, we recently ran a blog listing some of the most important attributes for you to look for in your colocation data center provider. We started by taking a deeper look at what exactly colocation is, and why it could be of benefit to you. If you missed it, definitely hop on over and check that blog out first, because it’s a great starting point for anyone considering colocation.

If you’re still reading at this point, we’ll assume that you’ve read that blog post by now. So, that means it’s time to pick up where we left off and detail some of the other important factors to consider when thinking about colo. Read on for the good stuff.


Here’s one that sounds obvious, right? We can hear you saying, “Of course I’m going to consider pricing of anything before I buy it.” Fair enough, dear reader. Spending less money always looks better on the bottom line than spending more money. But there’s more to business than playing not to lose, you have to spend money to make money and other such clichés.

Now then, let’s put aside rock bottom pricing and take a look at what different pricing levels really mean when it comes to colocation – what are you going to be getting for your dollar? It costs a lot of money to construct, manage and operate a data center of your own. Putting your servers in somebody else’s center will save you money. However, it’s vital that you understand that facility differences, service packages and quality of IT staff mean that all colocation sites are not created equal, so saving as much money as possible may not be the best strategy for everyone.

Continue reading Hallmarks of a Colocation Center Provider You Can Trust: Part II