The evolution of cloud services

17th October 2011

FSN writer Lesley Meall looks at the evolution of cloud computing and the latest developments from cloud vendors, and finds that clouds are not what they were.

It is the nature of the beast that any emerging technology will be in a state of flux, and cloud computing is no exception. In the beginnning, when the cloud was merely a euphemism for the internet, things were relatively simple and straghtforward. ‘The cloud’ term of reference applied to on-demand, pay-as-you go (or free) computing services that could be accessed by anybody with an internet connection and a browser. But as the cloud computing model has evolved, more technology companies have jumped on the bandwagon, more types of cloud service have emerged, and the whole soggy mess (and its myriad terms of reference) have become progressively more difficult to follow. 

Back when the cloud was about publicly available services, the technology industry seemed to be in agreement about how to best classify what was on offer. It opted for Software as a Service (SaaS), Infrastructure as a Service (IaaS) and Platform as a Service (PaaS). These groups aren’t too difficult to grasp, but if you are struggling, you might want to take a look on some previous FSN coverage explaining what these terms of reference mean, and how they operate, here, here, and here. Put simply, SaaS, PaaS and IaaS all describe ways of accessing data, applications and hardware remotely, through third-party providers; but things have become more complicated with the arrival of ‘hybrid clouds’, ‘private clouds’ and ‘personal clouds’. 

The cloud is still awash with public services, and the basic delivery model hasn’t changed. Using the internet and a browser (such as Chrome, Firefox, Safari or Internet Explorer) you can still access as much or as little as you want of a range of services that are physically located far away (in ‘the cloud’) on computers that belong to a third party provider. Think software applications (, computing power (Amazon Web Services), data storage (Dropbox, which FSN covered here), and the tools to develop web-based applications ( Try not to think about the fact that cloud service providers use the services of other cloud service providers that also use the services of other cloud service providers, and so on, or you’ll end up feeling as if you are lost in an Escher drawing

As all of this complexity can potentially deliver flexibility of access, economies of scale, and improved utilisation of resources, amongst other things, it is hardly surprising that businesses were keen to enjoy these benefits without putting their data on the same box as lots of other businesses, and/or outside the corporate firewall, so private clouds were born. Well, sort of. The very many providers of traditional managed computing services (in their very many forms) were also keen to exploit the benefits of the cloud (including the media hype), so they took the opportunity to re-position their offerings. It’s one of those chicken and egg situations, and it’s hard to be sure which came first. Either way, we are where we are today: faced with a range of private cloud offerings. 

This is where it gets a bit complicated. A private cloud can comprise just one computer server, containing software applications (such as software packages for accounting and CRM, or tax and audit) and storing the associated data, and all of this can be accessed (remotely) by people who have the necessary authorisation, internet access and a browser; but many private clouds are just a variation on the traditional data centre containing numerous boxes. Sometimes these boxes are owned by the business, operated by the business and provide computing services to the business. Sometimes the boxes and the software are owned or licensed by or to the business, operated and managed by a third party, and provide computing services to the business. Sometimes, the boxes are owned or leased by the service provider, operated by them, and used to provide computing services to just one business; sometimes it’s multiple businesses. But because they are not providing pay-as-you go services to all-comers, they are all described as private clouds. 

Then there are the hybrids, which mix public and private systems. Even if they want to, not all organisations can dump all of their existing hardware and start using cloud-based boxes, so some organisations make the transition gradually. As servers reach the end of their useful life, they are succeeded by third party resources that are located in the cloud – and sometimes these are located in the public cloud, and sometimes they are located in private clouds. Either way, they’re still called hybrids. Then there are organisations that see the public cloud as a way of augmenting their existing computing infrastructure, by using cloud services to quickly expand and contract, in response to fluctuations in demand – and these are hybrids too. 

All of which brings us to one of the more confusing incarnations of the cloud, the personal cloud. If you have read this far, you won’t be surprised to find that there is more than one variation on the theme, which involves providing you with remote access (using the internet and a browser) to the data held on your own computers. Proponents include Iomega and iTwin. With Iomegu you can create a personal cloud by connecting an Iomega network-enabled storage device  to other devices (such as a PC, Mac, iPhone and iPad) via the internet. Once you have done this you are able to invite up to 250 people (or individual devices) into your personal cloud, enabling all sorts of people to access and share all sorts of data, copying and sharing files directly between computers, as though they were all together on a local network. 

The approach iTwin takes to providing a personal cloud is different. It also uses the internet and browsers to provide people with remote access to the data held on their home or office-based systems, but it does it in a different way, by using USB sticks (without doing anything as risky as actually putting your data on them, of course).

Rather than controlling access just using logical security, such as passwords, iTwin adds a layer of physical security, by using pairs of USB sticks that contain encrypted keys that are unique to you. If you leave one of your two USB sticks plugged into the computer your data is on, you can then access it from any other internet-enabled device, as long as you have plugged the other USB key into this device. 

This is possible, because iTwin verifies the presence of the ‘paired’ USB keys, and after it has done this, it transfers your encrypted data onto its own servers (in the cloud) where you can then access it. When you’ve finished, any modified files can be saved back to the computer your data is stored on, and temporary files on iTwin’s server and the local machine are auomatically deleted. It’s fairly clear what’s cloud-like about both of these offerings, but it’s also fairly clear that they both take the initial concept of cloud computing and stretch it somewhat. This doesn’t make their services any less appealing, or useful, but it does make it difficult for many people – who might potentially use them – to figure out where these ‘clouds’ fit in amongst the very many other types. 

As if all of this did not make ‘the cloud’ complicated and confusing enough, it is also becoming increasingly difficult to identify the business benefits that separate and public and private cloud services – and not just because of the increase in hybrid set-ups. Some of the characteristics that used to differentiate public and private cloud offerings are now common to both.Even vendors that wrote the rules on the cloud with their first generation public cloud offerings (such as Amazon Web Services) are becoming as ‘elastic’ in their definitions of cloud services as they are in their provision. The bright lines that fleetingly separated public cloud services (and their providers) and private cloud services (and their providers), are fading, as vendors in both camps change the range and nature of their offerings. 

This is particularly noticeable in the Iaas (or data centre) space, where users have always had to do more ‘hands on’ system management than users of SaaS. From the get go, one of the big selling points of SaaS was its ability to relieve businesses of the need to manage and maintain the software applications they use. This abdication of responsibility was less noticeable with Iaas, where users have always needed to do a certain amount of ‘management’ themselves. But because organisations increasingly want to offload as much of the ‘grunt work’ as they can to somebody outside the organisation, public clouds are becoming more like private clouds and private clouds are becoming more like public clouds. 

As Amazon evolves for example, it is increasing the ‘hand holding’ and support it offers to users of its on-demand pay-as-you-go Amazon Web Services. Earlier this year, it extended its Premium Support Range, offering users of its servers and data centres more options, higher levels of service, and in the case of its top of the range ‘Platinum’ support, charging significantly more too. Bronze support starts at $49 per month, Silver at $100, and Gold (which offers 24/7/365 support and one-to-one phone support) starts at $400, while Platinum leaps up to $15,000 or 15 per cent of AWS usage (depending on which is higher) – which is beyond the reach of the small and medium businesses that cloud-computing has previously levelled the playing field for. 

The Platimum service looks a lot more like a traditional managed service than you might previously have expected (or been able to get from) AWS. It offers a first guaranteed response time of 15 minutes, accounts have a named technical account manager assigned to them, and tickets get  ‘white-glove’ routing that prioritises them over tickets that have already entered the queue at other levels. Platinum also provides help with infrastructure planning, and account managers will conduct performance reviews, report on metrics, collaborate on launches, and connect account holders with solutions architects. 

But this takes AWS into direct competition with organisations that offer traditional managed services and co-hosting arrangements (with more than one tenant on each box), and businesses will want to do a careful comparison between the various options – and not just because of the cost or the service levels on offer. A Service Level Agreement (SLA) with a managed service provider typically offers financial compensation in the event that something goes awry; providers of public cloud services typically offer ‘service credits’ against future payments, as FSN has previously highlighted. Meanwhile, providers of traditional managed services are also evolving. 

Many already use virtualisation software, so that private clouds can offer some of the same benefits as public clouds. Commoditised services are also emerging. Hosted Desktop UK, for example, offers a fully managed hosted service that provides scalable access to software ranging from general productivity applications to more specialised industry-specific tools (along with storage, support and so on). The target market is currently accountancy practices, but the business model could be expanded to include other sectors too – so the economies of scale are probably going to lead to many similar offerings from other service providers. At some point, it’ll probably become a discernible trend with its own special (and doubtless confusing) term of reference. What type of cloud this will be remains to be seen.