Hybrid cloud: how enterprises can build resources to suit their own needs
Author: Beatrice
By Jack Bedell-Pearce, CEO and Co-Founder of 4D Data Centres
With so many issues that can cause inefficiency, IT leaders need to ensure the right foundations are in place in order to optimise the management of hybrid cloud. Every environment is different and there is no one-size-fits-all cloud infrastructure. So how can organisations prepare and build resources that work for them?
Why is optimising hybrid cloud management important?
Hybrid cloud is more than just sharing workloads between the two major hyperscale cloud providers, Azure and AWS. It also encompasses other infrastructure environments such as on-premises servers, private clouds, and servers in colocation.
No one platform is necessarily better than another, but it is important to regularly evaluate them individually to make sure they are ticking the right boxes. Five essential areas platforms that need to monitor are performance (inc. compute, latency, bandwidth etc), reliability, resilience, security and cost efficiency. In addition to this, green credentials have recently become a sixth important factor, with companies realising colocation data centres and some hyperscalers are able to offer significant improvements in cooling efficiency and, in the case of colocation, high density cooling for High Performance Computer (HPC) systems.
Not all platforms are equal when evaluating these criteria, so it is important for companies to consider what to prioritise in their business when matching their workloads with the relevant platforms. Public cloud is very good at providing entry level services and scaling quickly for fast growing businesses, but for more mature companies (especially those with readily available capital and potentially legacy systems), a blend of public cloud, private cloud, and colocation may be a more cost efficient and reliable option. This is demonstrated in a whitepaper by Andreeson Horrowitz, which shows the financial cost of enterprise companies miscalculating the mix and discovering significant cost savings by repatriating servers back into data centres.
The right foundations and implementing good practice
In the same way you wouldn’t advise someone to put all their savings into one asset class, large companies should avoid being overly dependent on a single platform. Aside from the obvious downtime risk associated with a single point of failure, there is the potential risk of being trapped and unable to avoid price inflation if your sole IT platform is provided by a third-party vendor.
Once the right foundations are in place, enterprises need to become more organised and build IT resources through good practice. Examples of this include:
● Governance – how do you ensure the business is aware and being fulfilled/responsive to departmental needs (without them going off and just doing their own thing)?
● Security/Identity/Access Management – making sure that as services spread out, the right people have the right level of access. Data leaks can occur through poor basic hygiene and configuration.
● Stepping back and assessing how they’re using what is deployed; an example of this is Brandwatch doing front end visualisation in GCP (Google Cloud platform), as they had some good assets for their development team but the backend data was stored in colocation.
How can optimisation pitfalls be avoided/mitigated?
In order to minimise mistakes, enterprises should orchestrate across different businesses to overcome the ‘one pane of glass’ challenge for provisioning and delivery, be aware and in control of costs and recognise different approaches. The different potential costs of hyperscale cloud vs running your own vs colocation should also be considered, with the cost of the equipment etc. taken into account.
Additionally, monitoring and reporting of the end-to-end solution using the right tools for multicloud/hybrid use must be factored in. This will ensure accurate and consistent alerting as well as raising awareness on what is actually being deployed and where, removing assumptions of resilience.
Other areas to be aware of are the overlap or expansion of products and services, so as each provider continues to expand their product set, integration must be consistent and done at regular intervals to avoid being left behind. Integrating services and applications can also help with silos, but businesses must be careful of non-standardised interfaces to avoid future migration nightmares.
Once hybrid cloud management is optimised, what should CIOs do next?
Whilst CIOs might get close, it is unlikely that they will ever fully optimise their hybrid cloud setup. As with all technology, trends and advancements are happening regularly, so being up to date is not something businesses can ‘fit and forget’. Technologies will continue to evolve and part of the role of CIOs is to ensure they are not left behind and are tweaking their infrastructure accordingly and frequently.
Perfecting cloud services demands a commitment to agility and change. Trends are endemic to the cloud and will continue to evolve at speed as adoption increases. Tracking and unpacking trends will help your enterprise to open doors by leveraging the expertise and knowledge of the industry. As the world continues to embrace cloud services, these opportunities will be essential to sustained growth in 2022 and beyond.
