Business IT is being dragged down by the weight of the ‘old’. Whether it is outdated assumptions, virtualisation platforms designed for consolidation not agility, or simply neophobia, this isn’t just disruption, it’s a death-grip. Losing the weight requires fundamental change, and here’s why:
Flexible and agile is the ‘new normal’
Enterprise success today relies on speed, agility, security and availability, while always keeping costs under tight control.
But yesterday’s IT is not up to the job
Many enterprise IT platforms were built to solve yesterday’s problems, resulting in hard to manage systems with limited flexibility.
Cloud capabilities are wanted
Public cloud systems have the capabilities that business IT requires, but not all workloads are suited to running in the public cloud.
New IT infrastructures are available
Running cloud-like systems internally and combining them with public cloud resources is the way forward, but it requires new infrastructure platforms.
IT needs fresh ideas
The pressure on IT is mounting. Can you meet it before it becomes too great for today’s systems and processes? Can you meet it before something breaks – and breaks badly? How much longer can you put off real change – not just extension and adaptation, but fundamental change?
The demands are everywhere, and are coming faster than ever. Users and customers want new IT services or updates at the frequency they expect from the public cloud. This puts pressure on your IT people, processes and systems to be able to deliver new applications to new users, any time, anywhere. You probably recognise most of this and may well have begun addressing the issues. It’s likely that Agile software methods and DevOps have already found their way into your kit bag, even if they have not yet scaled across the entire IT delivery system.
Are your IT platforms ready?
Like most organisations you are almost certainly building private clouds to improve IT flexibility and responsiveness, mimicking the capabilities you expect from public clouds. The challenge is how to blend these two together to enable a converged and hybrid approach to service delivery.
Building the hybrid future
The question then is what core solutions will you build your long term hybrid cloud IT platform on? This is challenging when it is likely that some of your existing kit will be several years old. How will you go from ideas and pilots to scalable and repeatable operations?
It’s time for a cold, hard, readiness check. It’s time to get ready to change. It’s time for
an infrastructure rethink.
Clouds bring flexibility
The growth of public cloud usage within business IT over the past decade has been dramatic. While some of this has been as a substitution of applications previously run on internal IT systems, much has taken place because of the core capabilities inherent to public cloud offerings.
Public cloud has great features
In many ways public cloud has set the scene – and, more importantly, the expectations of users and customers – in regard to flexibility, responsiveness, scalability and cost. You need a VM now? You spin one up in a few minutes at most. You need extra resources to meet a surge? No problem, public cloud is on-demand. You want to pay for the resources from Opex, not Capex? Public cloud has the model for you.
However, while public cloud is excellent in many respects, it has disadvantages too. As a result it isn’t able to meet every enterprise need or expectation. For example you may have specific service level requirements for low latency, or regulatory limits on the geographical location of data.
Taking it private
There are also questions to ask around cost and control, where private cloud may provide better answers. One that is too often ignored relates to ‘sunk costs’ – once you have the datacentre infrastructure and skilled staff in place, adding an incremental cluster or scaling an existing one up should be cheaper inhouse. Then you’ve paid for it, so it’s yours, with no ongoing fees.
In addition, many organisations have applications, services and data that, for whatever reason, they want to directly manage and control. That requires skills and infrastructure, and once you have those, you want to get as much use and value out of them as possible.
The message is that public cloud will be part of the mix, but alongside cloudy platforms in the datacentre as part of a hybrid public/private approach.
Virtualisation is not enough
There is an old joke in which a traveller asks a local for directions. The local pauses a moment for thought, then replies, “If I wanted to get there, I wouldn’t start from here.”
So it is with private cloud. Many organisations have tried building flexible, scalable and responsive on-demand systems based on their existing virtualisation technologies. The problem is that while first generation virtualisation platforms worked very well to improve resource utilisation and provide some degree of flexibility, they weren’t designed to meet all of the demands expected of true cloud solutions today.
For example, they may have practical limitations or present excessive complexity in areas such as moving workloads between cloud and on-site resources, the portability of security policies and virtual machines, and policy driven governance in hybrid scenarios, to name but three.
Commercial models are vital too
It is also important to recognise that the legacy commercial models associated with these
platforms were not designed with cloud in mind, whether public, private or hybrid. They may feature highly constraining contract terms, for instance, or excessive software licensing and maintenance fees.
Starting from the wrong point brings challenges
Virtualisation platforms were never conceived for the IT world of today, where it is essential to be able to move workloads without interruption and with little advance notice. When those systems appeared a decade or more ago, the primary intention was to drive consolidation and operational efficiency.
Yes, these platforms have subsequently been enhanced to cope with the need for more scalability, flexibility and to give responsiveness to emulate the on-demand model. However, the truth is that they are starting from the wrong point – and perhaps more importantly, with the wrong philosophy.