The term ‘cloud’ has certainly drawn a lot of attention recently, to the point where it’s become a catch all buzzword, devoid of much substance. The definition has been vague and poorly communicated by most of the big names in the industry, as they define it from their point of view and service offering.
In a general IT context, cloud refers to computing infrastructure or services that are accessed ‘remotely’. This is a pretty broad definition, so I will focus on the two most common examples, which are:
SaaS (Software as a Service): this includes search engines and webmail provided by the likes of Google, Yahoo and Microsoft. These are applications that sit on remote servers and are accessible via a ‘thin client’, usually your web browser.
IaaS (Infrastructure as a Service); this includes Amazon AWS & Microsoft Azure. In this model, you are essentially leasing virtual servers, although there’s some variation on what’s possible.
In the SaaS model, the infrastructure and the application are hosted and managed by an external provider. You simply connect and use their services or application via a ‘thin client’. Your browser (the most common thin client) displays the output, but all the computation is done remotely on the provider’s infrastructure, which hosts the application.
In the IaaS model, you’re only paying for hardware infrastructure resources. What you then install on that hardware is entirely up to you. The main reason for choosing IaaS over SaaS for any given purpose is that the level of flexibility and control with IaaS is much greater.
The benefit of these cloud models is that you don’t need to purchase or maintain any hardware (servers), and in the case of SaaS, the software aspect is also handled on your behalf. For day-to-day operations, this means you don’t need to worry about upfront hardware purchases, hardware failure, hardware replacement at end-of-life, license purchases, software updates, security patching, backup systems, associated electricity consumption and the physical foot print it takes up in your office.
Another less common benefit that some organisations are now seeing is that their laptop purchase costs are going down. The cost reduction comes from purchasing cheap low spec laptops primarily aimed at thin client usage, such as the Google Chromebook range. These low-spec devices have become useful thanks to the huge growth in SaaS applications, reducing the need for locally installed software and thus removing the need for expensive high-spec laptops.
With all the apparent benefits of cloud solutions, you might wonder why it’s only now becoming prominent. One of the biggest reasons for this is that virtualisation software has only recently matured, even though it was conceptually born around the 1970’s. To be quite specific, VMware, arguably the market leader in this space, really moved things along with the release of their ESX product in 2001.
Virtualisation software allows you take a given amount of hardware resources and divide it into discrete logical units. For example, if you purchased one physical server, virtualisation software allows you to divide that into multiple servers, each with its own purpose (e.g., mail, file, application). This is desirable because it increases the utilisation of your hardware, which generally reduces your costs.
Consider an example where you need 4 servers to run a mail server, file server, application ‘A’ and application ‘B’. You have the following options;
This concept of sharing physical resources and increasing their utilisation is taking advantage of the fact that most of the time your hardware is underutilised. You don’t need as much physical hardware to get the same job done, which saves you money. Taking this idea further, it allows for economies of scale that were previously not possible.
Virtualisation made scaling much easier and the cloud industry experienced massive growth. Cloud vendors could build a data centre with a virtualised environment and then divide the hardware resources amongst clients depending on their individual usage needs. They could maximise the utilisation of their hardware and if they ever needed to increase capacity, it was relatively easy to install more hardware and add it to the aggregate pool of infrastructure they already had.
Scaling to the customers demands was also much easier. If they needed more CPU or RAM because of increased work load, it could be virtually allocated from the cloud pool. With this level of control becoming available at the software level, automation and productivity, consequently improved. These days you can sign up to a cloud provider and use their web console to specify your requirements and have access to a server in minutes. If you only needed it for a short period of time, no problem, simply remove the virtual server from your account when your done. You only pay for the period you used it. This has been an excellent usage model for software developers.
With economies of scale and lowered barriers to entry causing an increase in competition, it was inevitable that prices would drop. Relative to in-house data centres, cloud vendors have lower variable costs because of their ability to purchase in large quantities. Their fixed costs are spread over a large customer base, and because they specialise in their field, their relative efficiencies are much greater.
The industry has reached a level of maturity and a price point where running your own data centre makes economic sense in only the largest of organisations. When it comes to small and medium businesses, it no longer makes economic sense to run your own server infrastructure.
With that all being said, are there any downsides or limitations to the cloud? The answer is yes, absolutely. We could write a separate article all about that, but for the purpose of this article, we will focus on the biggest problem that most small to medium businesses will come across when adopting cloud solutions. Quite simply, the problem is bandwidth and latency.
The biggest consideration when implementing or moving your applications into the cloud is – do I have enough bandwidth? Having your infrastructure on-site means that you access it over your LAN, which is relatively very fast. However, once you move into the cloud, you’re accessing your applications over a WAN, which is typically your internet connection.
Most small to medium businesses can’t afford dedicated internet links and many still run on ADSL. The most common use case where this becomes an issue is in network file storage and access. This is particularly true for organisations that need to access and store large files, as having to send this data over your internet link to a cloud solution can be very slow and, in most cases, unworkable.
Until fibre connections (NBN!) become more prevalent and prices for dedicated business links come down, for most circumstances, we recommend local file storage, whether it be a NAS or a dedicated file server, depending on your size and needs.
The two best use cases of cloud computing for SME’s today are email and telephony, or VOIP, as it’s usually referred to, but there are actually many more and the list continues to grow.
Contact us today to find out what the cloud can do for you.