The Case For Cloud

With the ever onward march of technology, several cycles of virtualisation have passed over the past decade. In it’s self, this is inevitable, as we are always driven by the need to develop and look for the next – the latest and greatest.

First, there were mainframe computers run with remote ‘dumb’ terminals connected – applications (in reality programmes running on punch cards and spooled tape) ran centrally and were viewed remotely. Next, there was client / server computing – the traditional ‘pizza box’ environment where clients ran their own OS, and connected to a remote server running another single OS. Applications were often run remotely, but often also ran locally collecting shared information from the server. After that came the traditional virtualized environment where companies, departments and individuals consolidated physical servers into virtual machines – releasing wasted resources to make running IT more efficient, more green, and more cost effective. What came next…?

….Cloud Computing arrived.

So, what does cloud computing bring us? Essentially it means different things to different people, but before we define the benefits to individual categories – we need to understand what cloud computing really is? Provisioning a cloud infrastructure can be done in many ways, using many different tools and platforms, but when all is said and done – the end result is the same. Essentially – it is virtualization cut loose – it is virtualization without constraints of physical resource. A true cloud can host applications anywhere, anytime – irrespective of global location, underlying hardware, or implementation method. Of course there are constraints to this freedom – security and legislative considerations among others that must be accounted for in a cloud implementation, but I consider cloud computing to simply be this: Virtualization 2.0.

So what, I hear you ask. Why do I want all this freedom when I have my private virtualization solution installed and working very nicely – thank you very much – what can this cloud do for me? To understand this, we need to look back at what has gone before, to realise the potential of what we never knew we could have until now.

Why the move from mainframes to client / server? Mainframes were very expensive, and known only to the chosen few – beyond the reach of mere mortals. As technology developed, PC accessibility increased and the likes of Windows, Mac OS and Linux came were born. These and their server siblings played nicely, but were not always the most effective way for resources (human and mechanical) to be used. Waste became prevalent – to the point where a new solution was needed. Enter virtualization with VMware in the market lead, freeing up wasted resource and consolidating servers and management workflows to produce the highly available and resource effective platforms we use and enjoy in production today.

However, this promise of well utilised resource and management is all very well, but the cost of the infrastructure for a virtual solution can be higher than the usual pizza box equivalent – large capacity hosts, networking and shared storage. Pizza box environments whilst often cheaper suffer too – from provisioning pains – think of physical datacentre space of all those servers and additional cabling – all these cost money too. The over-riding benefit of virtualization is still availability, redundancy and resource efficiency which is why we are moving in the direction we are, but what if we could provision infrastructure whilst addressing the drawbacks of the previous generations, using the benefits of virtualization. Enter cloud computing – ‘virtualization 2.0’. Using cloud technology, customers can quickly provision individual server, grouped virtual applications or even whole infrastructures simply and quickly without the high costs of purchasing infrastructure or the provisioning pain of physical servers.

When talking about cloud computing , new terminology comes into play. Terms such as IaaS, SaaS and DBaaS begin to describe the cross over between technology and methodology in cloud computing, for cloud computing really describes a ‘service’, rather than more traditional computing environment. Cloud Service Providers give us all the ability to provision our virtual servers, applications or infrastructure without worrying about how or what hardware to use, typically accessed through web portals for ease of access. These services can be described as:

  • IaaS (Infrastructure as a Service)
  • Saas (Software as a Service)
  • DBaaS (Database as a Service)
  • etc
  • Etc

These services are typically offered on 3 models:

  • Allocated – where a fixed amount of resource is guaranteed to the customer, and the cost remains the same per month. (Good for those who need either to know they have the resource or to know the spend per month).
  • Reserved – where resources are allocation per customer but there is the ability to ‘burst’ over the reservation into a larger pool, where additional charges apply. (Good to reduce monthly cost but with flexibility of peaks and troughs in resource usage).
  • PAYG – ‘Pay as you Go’ – exactly what it says on the tin, where resources monitored and are paid for each month based on consumption. (Good because it provides the most flexibility – if you don’t use any resource in a month, there is no bill!).

So, we have discussed the benefits of the cloud,but in what way can we as consumers use these new service offerings? Here are a couple of useage scenarios with possible cloud applications.

Case Study One – IT Department DR Scenario.

The IT Director of BigCorp has tasked the VM team with saving money on the annual capital expenditure budget, whilst at the same time refreshing the new offsite DR provision for BigCorp. How to achieve this?

  • Option 1: Traditional Virtualization. Possible steps:
    • Size the environment to be covered by the DR requirement.
    • Investigate hardware platforms, check compatibility and find a physical host locations.
    • Negotiate rental of the physical location (with potential long-term contracts for best cost-effectiveness).
    • Purchase new hardware, and install it at the new site.
    • Configure the virtual environment.
    • Migrate data to the remote environment.
    • Test inter-site failover.
  • Option 2: Cloud Virtualization. Possible steps:
    • Size the environment to be covered by the DR requirement.
    • Purchase resource from a cloud provider, either on allocation or ‘pay-as-you-go’ basis (depending on requirements).
    • Provision new virtual infrastructure using existing virtualization tools in the cloud.
    • Migrate data to the remote cloud environment.
    • Test failover provision.

Result? The cloud option greatly simplifies provisioning the new environment, whilst meeting the requirement save CapEx from the budget.

(Time Saved) + (Money Saved) = Double Win for Cloud Solution!

Case Study Two – Researcher at Smallville University.

As part of a research grant, a researcher at a small university is doing work on computer modelling for a paper. As the work progresses, more and more data is generated but there is no resource to process the data.

  • Option 1: Traditional Virtualization. Possible steps:
    • Apply to the university for funds / computing resource.
    • Wait for provisioning of the resource (sometimes weeks, and often not to specification), or engage in physical server build.
    • Gain access to the VM, often with restrictions applied.
    • Build compute processing application on provided provision.
    • Deploy code for processing.
  • Option 2: Cloud Virtualization. Possible steps:
    • Browse to cloud provider on the Internet.
    • Sign-up for an account in 5 minutes.
    • Provision a server directly into your cloud, to your specifications.
    • Gain access to the VM with no restrictions applied.
    • Build compute processing application direct to cloud.
    • Deploy code for processing.

Result? The cloud solution can be provisioned quickly and easily, with billing direct to the researcher if desired, saving again time and money by bypassing central traditional deployment leadtimes and budgetting.

(Instant Access) + (Greater Flexibility) = Double Win fore Cloud Solution!

Conclusion.

So – as we have seen, there are some serious benefits to cloud computing models over traditional virtualization architectures. We will write more on these benefits, the cloud technologies available and future developments as time passes. One important thing to point out is I’m not suggesting we all replace or envelop our existing platforms into cloud solutions. As with everything, there is no one solution for every requirement, but adding some cloud into the mix is taking the first step towards releasing the benefits of the new technology.

In my opinion, cloud is here to stay, and will only increase in popularity as technologies mature. There is something for everyone in the cloud market – from open source code to enterprise vendors with fully-featured solutions. So maybe it’s time to dip your toe into the cloud?

Jeremy loves all things technology! Has been in IT for years, loves Macs (but doesn't preach to others about their virtues), loves virtualization (and does shout about it's virtues), and sometimes skis, bikes and directs amateur plays!

Leave a Reply

Your email address will not be published. Required fields are marked *