From the very first days of public Cloud computing, there’s been an ongoing debate about which is superior for hosting a website/application: Physical servers or the public Clouds.
The truth is that there’s no single correct answer that applies to all customers and in all situations. There are circumstances in which physical servers offer the best solution, whether based on performance or cost, as well as situations in which the opposite is true — and the public Cloud is the clear winner due to its various advantages (scalability, zero-maintenance, etc.).
The real problem in this debate is that when a customer approaches a Public Cloud provider, that provider will promise him (and yes, all other customers that inquire) that the Public Cloud is his ideal solution. When that same customer turns to a Colocation provider, that provider also assures him that his own approach is best. And to be honest, it’s hard to blame the providers for working hard to sell their services and not someone else’s… but that’s why we are here.
Because we’ve done the research and have been certified by each of the leading Public Cloud provider, and after over a decade of experience in the Colocation sector where we launched hundreds of projects with physical servers, we are uniquely qualified to take a step back, and carefully examine the realm of hosting options. We understand clearly the benefits and drawbacks of each Public Cloud service provider, just as we know the benefits and drawbacks of Colocation. As such, we at XGlobe can produce a well-defined working plan that we call Cloud Consolidation, incorporating the best of these two environments. This ability allows us to maximize performance of the unified system, and to substantially and measurably save on monthly expenses.
Here is a short example of the effectiveness of
Take, for example, a website that sells tickets to performances. On average, about 10,000 users visit the site each day. Now assume that the site experiences traffic peaks when extremely popular shows are advertised — and at those points, 300,000 to 400,000 visitors a day arrive at the site looking for tickets.
According to these figures, we know that the sites owner needs to prepare an infrastructure for 10,000 people on a normal day — about 70% of the time. But in addition, he has to be ready for those peaks that will occur 30% of the time, bringing in 30 to 40 times the usual traffic. The right solution for the “normal” traffic is an implementation of physical servers or virtual ones, at a fixed cost. After all, the site owner knows how much traffic to expect, requiring only a fairly standard implementation that most hosting providers can readily offer.
But for those peak times, he certainly needs a scalable infrastructure that can quickly, smoothly and without limit handle the load.
Of course, this client can start out buying physical or virtual hardware in the cloud, but he'd find himself paying thousands of percent too much simply to cover that 30% of the month. In short, it’s a tremendous waste of money.
The recommended service for dealing with these peaks is Autoscaling from AWS, a “Pay-as-you-go” service. This service automatically adds instances that can support these extraordinary traffic levels, and at the end of the month the client only pays for actual use.
There is, however, a significant drawback: there should be a reasonable cost under normal conditions where only 10,000 visitors arrive; here, for that extra scalability, he is paying a lot more than he would for a standard implementation of physical or virtual servers.
It is in precisely these types of situations that XGlobe steps in to offer its clients more creative and practical solutions that blend multiple hosting providers, saving money and benefiting from performance along the way.