Everyone is talking about the hybrid cloud, but what is it? Is it right for your business? And if so, what should you consider when shifting to a hybrid cloud infrastructure?
First, let’s consider recent trends in cloud computing more generally. Cloud computing has its roots in the 1950s, with the invention of mainframe computing, but it only really took off in the 1990s, when telecommunications companies were able to start offering virtualised private network connections.
This opened up the possibility of public clouds, where a third party such as Amazon or Google – or, commonly, a combination of such third parties – owns the physical hardware supporting data storage and applications, and this infrastructure is shared by multiple clients. Cloud-based server hosting, webmail, and online office applications like Office 365 are some of the most common forms of public cloud service today.
Private clouds, where hosted services are provided through a proprietary computing architecture owned and managed by the individual client, soon followed.
There are different benefits on offer with both models. Public clouds give organisations the scalability and flexibility of depending on hardware managed by a third-party provider, while private clouds offer the superior control – and, potentially, security – that comes from being the sole owner and manager of a cloud environment.
Hybrid clouds then grew out of a desire to marry the different benefits on offer with both public and private clouds. Hybrid clouds combine dedicated, localised servers and infrastructure with cloud servers and storage, both public and private, on the same network. By mixing and matching various combinations of these different network approaches, from localised datacentres to cloud services hosted by third parties, and even managed private clouds, organisations can build highly flexible environments to meet their IT infrastructure requirements.
For example, it might be important to host highly sensitive, business-critical applications and data, such as customer financial data and processes, on premise, while less critical services, such as office programs, are hosted by a public cloud provider. As such, hybrid clouds are the most flexible and adaptable types of cloud infrastructure – which is why they are increasingly popular.
In fact, according to Gartner, by 2020, 90% of organisations will adopt ‘hybrid infrastructure management’ – meaning a combination of cloud, hosting and traditional infrastructure services. Whether organisations are primarily focused on reducing costs, increasing scalability, shoring up security and compliance, adding new services or enhancing existing ones, creating a hybrid cloud infrastructure is, in many cases, the best fit answer.
However, all this can seem a little overwhelming if your business is still at the beginning of its hybrid cloud journey. Given all of these different considerations, how can you ensure that you create a hybrid environment that is a good fit for your needs?
There are five main areas that I think you need to consider:
Datacentres have hefty running costs in terms of power, cooling and maintenance – which means, in turn, that storing data locally can be expensive, sometimes prohibitively so. Paying for a cloud-based ‘as-a-service’ model whereby you only pay for the storage or compute power you require can be an obvious cost-saving alternative.
However, it’s not always that simple. Depending on the infrastructure you currently have in place and the outlay that will be involved in creating a range of virtualisation and infrastructure set-ups, a new cloud environment can be prohibitive in itself. This is why it’s important to have a value-based approach to analysing the cost and efficiency benefits of public and private cloud in different parts of your proposed infrastructure.
Don’t forget that datacentres date and require significant upkeep and upgrading at periodic intervals, so your cost and efficiency analysis should be long-term in outlook, considering what additional one-off costs may come into play in one, two, five years and so on.
It’s important to remember that the reality of cloud storage provided by a third party is that an external organisation is ultimately responsible for the integrity of your data. In other words, you are extending your cybersecurity chain to involve that third party.
Key areas to consider include data redundancy, what your provider’s uptime record is, how they respond to outages and the Service Level Agreements (SLAs) that you have in place. If you are working with multiple different cloud providers –something which can happen in many hybrid cloud models – this becomes even more complex.
When it comes to integrating security controls such as authentication and authorisation across a hybrid cloud environment, you have two broad options. You can either replicate controls across the different clouds that make up your hybrid infrastructure, synchronising the data across them, or you can deploy an identity management service that delivers a single security service across the whole environment. There are cost and simplicity of management benefits to weigh up between the two options.
Depending on the nature of the data generated and processed by your organisation, there will be strict laws and regulatory requirements in place governing where it can be stored and how it can be treated – and the chains of responsibility for these demands will necessarily extend to the cloud providers you work with. For example, any medical data must be stored in its country of origin, and this will govern which public cloud providers you can work with – you need to check where their datacentres are based. If you are subject to PCI DSS, then your providers need to meet those restrictions too.
Whatever the final design of your hybrid environment, it is vital to maintain seamless access to both cloud-stored data and that which remains in localised datacentres. This is where managed network operations can be invaluable, ensuring that visibility, monitoring and incident response is carried out at the same high level across all areas of the infrastructure.
As referred to in terms of cost and efficiency, it is vital to have a medium to long-term approach to any hybrid cloud investment – which includes planning for worst-case scenarios such as damage to a specific datacentre. It is important to map what impact such a problem will have on services running either in the cloud or through localised datacentres, and spread critical services and applications accordingly.
Similarly, it is important to consider your organisation’s overall roadmap for IT infrastructure investment, and how the hybrid cloud could fit into this. For example, if your organisation is planning to upgrade to a localised datacentre in the near future, then adopting a hybrid cloud model might allow for phased migration of data and a specific path that wouldn’t have otherwise been possible.
Hybrid cloud environments are here to stay. By blending different elements of public and private, on premise and externally hosted services and data storage, organisations can likewise blend the benefits on offer from each approach. However, maximising the opportunities of the hybrid cloud depends on understanding the precise needs of your own organisation, and the pros and cons that need to be weighed up across a range of areas, from security and compliance to cost and efficiency.
Where data is particularly sensitive, or even governed by specific regulations as in the case of healthcare data for example, a hybrid cloud model that incorporate significant use of localised services is likely to be preferable. However, if you’re planning a significant revamp of existing IT infrastructure, the use of either public or private cloud services might be key to achieving the flexibility you need.