In our last blog post we talked about what cloud is, the layers of cloud, and how it all fits together. But we never really looked at exactly what problems cloud itself solves, and how organisations can get started with cloud adoption. In this post we’ll talk about some general organisational challenges it can help resolve, as well as looking at some specific areas in which cloud technologies offer a compelling solution.
The Modern Mobile Workforce
One of the core characteristics of the modern workforce is that of mobility. The ability to work from anywhere, on any device. In recent years as IT infrastructure has tried to catch up with this change on work ideology, it has become increasingly complicated and expensive to build enterprise grade mobile capable systems. Datacentre internet links need to be made redundant, datacentres need secondary disaster recovery locations, security appliances need to be deployed, enterprise identity management needs to be integrated, mobile device management needs to be deployed, and the list goes on and on for things that need to be configured, deployed and managed. These can all become significant blockers when it comes to enabling the mobile workforce and creates the perception that IT is in the way of progress.
But one of the core tenets of the public cloud is that it is accessible from anywhere. When you pick a public cloud based solution, none of the infrastructure engineering challenges are problems that you the consumer need to solve, generally speaking, the platform has taken care of all of them for you. This is especially true of SaaS based solutions.
Taking Office 365 as an example of this. Microsoft provides Office 365 as a highly available, always on, accessible anywhere platform. When businesses look at consuming Office 365, they don’t look at how they engineer a mail platform, or how they configure their datacentre aware load balancers. They look at what policies need to be enabled to ensure they meet their regulatory compliance requirements, they look at how they can use Office 365 and Azure Information Protection can better secure invaluable business IP, they look at how global workflows can be enhanced with Skype video conferencing. Adopting Office 365 becomes a journey in enabling better business outcomes immediately by consuming a SaaS platform rather than building infrastructure that will eventually provide services.
Scale to Cloud
Another challenge that many modern businesses have to manage is the challenge of scaling their infrastructure for very high performance during very short peak seasons. In the traditional world of infrastructure this meant having a lot of spare capacity sitting about to soak up the massive performance spike that happens for 3 weeks in a year. Often times this infrastructure was repurposed for the bulk of the year as development and tests systems, or some other purpose was found for it so that it wasn’t just completely idle outside of peak season. But fundamentally, when designing and building on premise systems, organisations had to purchase enough hardware to sufficiently serve their peaks, however short they may be.
When deploying application systems onto public cloud this design paradigm can be thrown in the bin. Public cloud providers operate at a hyperscale that doesn’t exist in many on premise deployments. When applications get deployed to the cloud, scaling to cater for the peaks of the application is merely a matter of requested more VM or application instances be provisioned by the cloud provider and placed behind the load balancer. The capacity management is a challenge for the cloud provider to manage. As a resource consumer, the business also only has to pay for the peak workload during peak season itself, during the off peak season the business can operate with its standard steady state workload with its regular consumption cost. For those organisations with very high peaks this can represent a significant cost saving for delivering consistent high performance applications.
The Cloud Safety net
Following on the idea of leveraging the scale of public cloud to solve traditional infrastructure challenges, another common challenge that can be solved by looking at public cloud solutions is that of Disaster Recovery. Capacity management for elastic applications workloads and standby disaster recovery capability share a great many similarities. Both are required to deliver consistent high performance applications, but both often require significant capital and operational expenditure to achieve. In the traditional view, building disaster recovery meant having a second datacentre, with a full hardware replica of the production datacentre, physical servers, switches, SANs, power supplies, network links, all running in standby in case the primary datacentre failed. Some organisations took an active/active approach to disaster recovery design and built application systems that could split their workload between two active datacentres to try minimise the wasted standby capacity, but ultimately, spare capacity had to be sitting around and available in some form or another to cater for running production workloads in a secondary location in case the primary failed.
Azure, using its hyperscale Infrastructure as a Service platform offers an interesting solution to this problem: use Azure as your disaster recovery site. Azure Site Recovery offers a VM based replication technology that can replicate your VMs into Azure, for a per VM protection fee and the data storage costs only. You don’t need to pay for the compute capacity matching your production site until you initiate a disaster recovery event. Even then, you only pay for the compute running in Azure in a disaster while you are actively running in Azure, as soon as you fail your services back to your primary datacentre the cost ceases and you are back to only paying the protection fee.
This approach has the advantage of not requiring a huge capital outlay to prepare a disaster recovery site. Only an operational expense is required each month for your VM protection fees. A secondary benefit is that it also simplifies capacity planning and expansion, you only need to manage planning at the primary site, expanding the DR capability is simply a matter of enabling more VMs to be replicated to the cloud., and doesn’t require you to purchase a full set of matching hardware for DR every time you expand production.
This is just a small peak at the numerous problems that adopting cloud technology can solve for your organisation. For a broader look, or maybe a very specific look at your organisation, get in contact with your Dilignet account manager.