How to get hybrid cloud to live up to its promiseBlog
By Daniel Greengarten | Group Senior Vice-President, Sales, Managed and Cloud Services, Dimension Data and Kevin Leahy | Group General Manager – Data Centres
For established businesses, and those born in the cloud, hybrid cloud promises many benefits.
It allows you to develop and deploy applications, and get value from them, faster. It presents an opportunity to refactor legacy applications and place workloads on whatever platform is most appropriate.
It promises infinite scalability on demand, and the possibility of reducing capital expenditure, moving more towards consumption-based pricing models to match your business volumes.
In all, it promises to be the platform for digital business you’ve been looking for – one that moves as quickly (or as slowly) as your business needs to.
And, done right, hybrid cloud can be all these things.
Has it delivered what you wanted?
Many companies are facing serious challenges getting hybrid cloud to deliver on its promise. Companies come to us and say:
‘I’ve got workloads all over the place’
‘I’ve got no visibility of where everything is’
‘My compliance risk has increased’
‘The cost savings I was hoping for haven’t materialised’
‘It’s becoming very hard to manage it all in a joined-up way’
If you can relate to any of these, then you need to master the critical success factors of hybrid cloud…
The 6 critical success factors of hybrid cloud
You need to define a workload placement strategy that will give you the best balance of performance and cost, while meeting all your various governance, compliance and risk management requirements.
Then you need to assess your current ability to execute your strategy, focusing particularly on getting access to a service integration capability that will allow you to manage all the components of your hybrid cloud environment in a consistent and automated way.
You need to get visibility of your whole estate (actually not just hybrid cloud, but your entire hybrid IT environment), then segment it at every level, from macro domain to individual container, in order to limit threats.
You need security to be automated and integrated with the rest of your cloud management, so that controls are applied as quickly as compute, storage, or network resources are spun up.
You need a platform-independent management capability which abstracts away provider-specific complexities to effectively de-couple workloads from the underlying infrastructure. And you need thoroughly tested, scalable reference architectures to enable your workloads to be deployed with confidence on any cloud you wish.
Rather than trying to build a hybrid cloud platform yourself, find a preferred supplier who offers enterprise grade public cloud alongside private cloud on their own platform – and can integrate with your own on-premise or hosted private clouds, and hyperscale public cloud providers.
Workload automation is critical. In order to manage workloads automatically across multiple clouds, you need an abstraction capability with a library of common micro-services that all workloads require.
You also need a business portal that gives you end to end visibility of your clouds so you can see where everything is, troubleshoot quickly if necessary, and track trends to spot opportunities for further automation.
Lack of skills is a common challenge, and many companies find it easier to access the skills they need on tap through a partner than to attract and retain them themselves. You also need to decide how much of your hybrid cloud you want to manage yourself versus how much of it you want provided as a managed service.
Cloud networking independence is key. Rather than have different ways of connecting to each cloud provider, and your own on-premise private clouds, you need the capability to connect to all your clouds in a consistent way so you get an end to end view of application performance.
You may well need to re-architect and upgrade your own data centre network so that the portions of your workloads residing there perform as well as those in the cloud. Not only will you need bigger pipes and faster switches, but you’ll need to software-define your data centre network, as well as update your management processes and skill sets.
In any hybrid cloud strategy, re-architecting the WAN is critical to avoid performance issues and cost escalation. Most WANs were architected pre-cloud and data traffic patterns have changed – routing traffic via your own data centre may cause performance issues and incur extra carrier costs, but placing Internet breakout nearer the branch may expose new attack surfaces. Proximity to the cloud via low latency connection is necessary.
Hybrid cloud strategy even extends to the access network. Wi-fi access points and access layer infrastructure must be fast and smart. You need the ability to prioritise by data traffic type to ensure a good user experience.
5. Consumption models
Cloud has given the enterprise market an appetite for consumption based pricing as a way of avoiding capital expenditure and matching costs to business volumes. But consumption-based pricing isn’t right for everyone, so look for the ability to blend predictable costs with variable ones in whatever way suits your business best.
6. Business continuity
Business resilience starts at the strategy stage, so do a business impact and risk analysis, and design an architecture that’s sufficiently resilient where it matters, but isn’t unnecessarily over-engineered.
Once operational, make sure that all elements of your hybrid cloud are backed up, wherever they are, and that both applications and data can be recovered in the right order, preserving all relational dependencies between them.
If you’d like a more detailed account of all these considerations, and the answers to the challenges they raise, read our Best Practice Guide: Mastering the 6 Critical Success Factors of Hybrid Cloud.