In many ways, there wouldn’t be a cloud computing business without Amazon. When the company made its first Infrastructure-as-a-Service (IaaS) offering available to customers, it set in course a technology revolution that will permanently shake up the way that businesses work.
That’s a slight simplification of the facts. The concept of cloud computing has been around for some time. The phrase was coined in the 1990s (there’s a bit of a dispute about the first person to use it) and the concept of computing services being delivered as a utility was first suggested in the 1960s in Douglas Parkhill’s book The Challenge of the Computer Utility.
During the early part of this century, there were several attempts to introduce a rudimentary form of cloud computing. These went by various names: ASP, grid computing or utility computing before cloud became adopted as the accepted nomenclature.
Amazon was kicking at an open door and it’s also fair to point out that the emergence of Salesforce as a viable CRM option had a lot to do with the acceptance of cloud.
Amazon’s smart idea was to use the expertise that the company had acquired in running the infrastructure used in its massive warehouse operation. As Werner Vogels, AWS's CTO put it, "From experience we knew that the cost of maintaining a reliable, scalable infrastructure in a traditional multi-datacentre model could be as high as 70 percent, both in time and effort, and requires significant investment of intellectual capital to sustain over a longer period of time. The initial thinking was to deliver services that could reduce that cost to 30 percent or less. We were also keenly aware that compute utilisation in most cases, enterprise as well as startups, is dramatically low (less than 20 percent and often even lower than 10 percent) and is often subject to significant periodicity." From this premise, the company built AWS to where it is today.
To services its customers, the company has constructed a series of datacentres and, crucially, built up its infrastructure using its own servers. This means it can customise its infrastructure to meet its own intensive demands, and take advantage of huge economies of scale and buy components incredibly cheaply. This means it’s not using the likes of HP, IBM or Dell to supply its servers but working on constructing its own hardware – allowing the company to pass on the costs to customers and, in a cut-throat business, Amazon is generally among the cheapest.
There’s much talk of Amazon’s cloud services but in reality the company offers a broad range. Its two best known offerings are its Simple Storage Solution (S3) and its Elastic Compute Cloud (EC2). The former, as its name implies, is Amazon’s cloud-based storage (the company also offers a cheaper storage service called Glacier) while the latter is its IaaS offering.
Beyond these, there’s a product called Amazon Map Reduce (EMR) that allows businesses to process and analyse vast amounts of data; a virtual cloud offering called VPC that allows for the connection of a number of EC2 instances; Amazon DynamoDB, a database-as-a-service product, and Elastic Beanstalk, the company’s take on Platform-as-a-Service (PaaS).
There are three ways to buy Amazon cloud services. A customer can opt for pay-as-you-go which, as the name suggests, just requires the production of a credit card. In just a few minutes, a user can be working on an Amazon instance. In addtion, there is a Spot Instance option where potential customers can bid for unused Amazon capacity. Customers set the maximum price they're prepared to pay and if their bid exceeds the current spot price then their order is fulfilled.
The other way is via the Reserved Instance option: this is where a customer pays a one-off fee and that entitles him or her to a discount on Amazon services. In September 2012, the company introduced a Reserved Instances Marketplace, a way for organisations to resell its unused Reserved Instances.
Companies that want to go down the Amazon route should make use of the cloud calculator, a way for assessing what the service is likely to cost. The company makes it plain that this is a very rough guide to costs and people who use it have remarked that the final bill can be different. The other thing to be wary of is that ease of use does make it very simple to call up new services and bills can rack up very quickly.
Amazon has been amazingly successful at attracting customers to its cloud service. It’s not clear how many as the company is notoriously secretive about the number and its revenue (AWS is not accounted separately but is part of the Amazon empire). However, all estimates suggest the company is way ahead of its nearest rivals – with some suggesting it has 75 per cent of the cloud market sewn up.
But there are some black marks against its name. It’s a very no-frills service, offering little in the way of hand-holding but the company doesn’t claim to be anything different. There’s the issue of its datacentre reliability, too.
The company has been hit by several datacentre outages, notably one on the east coast of the US in April 2011 that left several customers without services for up to three days. There have been a couple of other smaller outages since, serving to remind users of the importance of having a good cloud backup plan.
There’s every indication Amazon is set to hold sway in the cloud market for some time to come – despite the emergence of a variety of new entrants to the arena. It’s still the first port of call for many organisations when it comes to opting for cloud.
Update: This article is an updated version of a previous article
