Blog 2: ‘Cloud-only’ is dead. Long live workload strategies
In our last blog, Post-mortem of a cloud-only approach, we discussed the difference between cloud-only and cloud-first strategies. And we joined Organisation X as they took their cloud-only journey.
Sadly, it wasn’t a happy outcome for Org X. Still, their story illustrates the dangers of following a top-down directive that compels an organisation to implement cloud solutions only – to all services and systems – and frequently at scale. Regardless of perfectly viable alternatives or more considered approaches.
However, that doesn’t mean the cloud-first approach is necessarily the answer either.
Cloud vendors have been fast to lay claim to ownership of all the associated cloud-first activities, a misalignment which ignores the deep experience of those who manage traditional cloud platforms, and the savings and innovation their services can deliver.
What is most important, though, is to maintain a tight focus on your cloud transformation goals. And for most organisations, that includes achieving greater business value and insights while controlling – or preferably reducing – costs.
This is where the importance of understanding your organisation’s workloads comes in.
Why a well-planned workload strategy is critical to achieving organisational cloud goals
First, let’s look at what workloads are.
A workload is an application, capability, service, or a specified amount of work that uses your computing resources. Logically, a light workload uses few resources, whereas a heavy workload demands significant amounts of computing resources. A cloud workload is identical in concept but relies on your user-pays cloud-based resources (i.e., computing, memory power, networking).
Currently, the most prevalent workloads across the cloud – public, private and hybrid – are databases, analytics, and web and content hosting. Other common cloud workloads include:
- Edge solution support
- IoT/data streaming
- CRM
- Industry market solutions
- Software development/DevOps
- Mobile app/device backend
- Legacy/app migration to containers
- AI/cognitive/machine learning
- Backup/disaster recovery
- Security
Next, what is a workload strategy?
In short, a workload strategy focuses on your organisation’s desired outcomes and existing application/data portfolio and then helps you identify the ideal (cost-effective yet high-performance) platform/location for those portfolio elements.
The Gartner article Design IT Infrastructure Strategies Flexible Enough for the Unknown suggests that “I&O leaders need a more flexible strategy to deploy any workload, anytime, anywhere.”
And we agree. Gone are the days of relying on one cloud (private or public) or on-premises platform to meet all business and workload requirements. As I&O leaders take an increasingly farsighted and strategic view of workloads, it’s widely expected the result will be the broad adoption of a hybrid model of public and private cloud, SaaS, and PaaS to provide organisations with improved business value and insights.
And despite repeated rumours of its ‘death-by-cloud’, continued investment in on-premises IT infrastructure will remain essential to support some traditional applications’ workloads.
Forrester’s Are You Getting The Most From Your Hybrid Cloud? leadership paper reports that “…IT leaders are tasked with pairing application workloads to the right combinations of cloud and traditional infrastructure to improve application workload effectiveness.”
Based on their research, Forrester says that ‘more than a quarter of (503) surveyed IT leaders consider high performance to be a top-three criterion when deciding where to deploy application workloads” and concluded that “…no individual infrastructure option meets the needs of all application workloads.”
However, spreading workloads comes at a cost. A hybrid model typically attracts a higher price tag than achieving the same outcome fully on-premises or entirely in the cloud. But the business outcomes that a hybrid model will deliver are more significant.
Circling back to Gartner’s designing-for-the-unknown article, they recommend that you determine your workload placement strategies based on business values – namely:
- Risk – if this workload fails, what’s the risk to your business and customers?
- Performance – does the workload require high performance to support queries, business intelligence, analytics and intense research operations?
- Regulatory requirements – are you subject to data residency and local sovereignty laws?
When should you put your workload strategy into play?
Today, a comprehensive analysis of your organisation’s workload to determine a workload strategy is considered a vital first step in successful cloud adoption.
A workload analysis helps determine timing and IT needs, and ensures that your chosen migration strategy aligns with your business goals (i.e., achieving more business value and insights while controlling and reducing costs).
Gartner recommends that these 7 workloads, in particular, should be moved to the cloud: Mobility, collaboration and content management, video conferencing, virtual desktops and remote workstation management, scale-out applications, disaster recovery and business continuity solutions.
You’ll note that a common requirement for most of the workloads Gartner suggests is demand variability and the performance burdens they place on networking bandwidth and computing resources. However, it’s only by completing your own workload analysis that you will be able to determine the big 3: Pain (how difficult life and business will be if you don’t shift it), potential payoff (the value to your users – both employees and customers), and priorities.
The workload strategy payoff
With a workload strategy in hand, you have the opportunity and knowledge to design a hybrid cloud model consisting of multiple cloud solutions. This will allow your IT teams to choose the cloud service best aligned to run specific cloud workloads to meet the requirements of the application and the organisation.
And if you leverage hyperscale cloud services strategically and are smart about your cloud instances and use of data replication to avoid expensive two-way data traffic, you can reduce and mitigate costs, increase availability, manage data privacy and compliance more precisely, and have greater geographic control over where your cloud workloads are run. Which should tick most of your business goal boxes.
Automating and simplifying workload management
Critical to controlling workload costs in a hybrid cloud model is the ability to manage data.
With a solution like ONTAP (NetApp’s data management software) at the heart of a hybrid multi-cloud model, it is easy to add new workloads from Oracle, SAP, SQL Server, virtual desktops and servers, and more, and provision storage for them in less than 10 minutes.
ONTAP allows you to move your data seamlessly between architectures to place it in the optimal environment for performance, capacity, and cost efficiency – and even automate where your data goes (including replications). And with real-time cost analytics, you can see which workloads are running where and the costs incurred by your chosen cloud providers – so it’s easier to make fast, well-informed decisions.
What else is important?
Regardless of which strategy you opt for (although we’re firmly in the workload strategy camp), it’s critical that you take your people with you. They will always be key to driving value.
For many, transition to a new technology environment or approach is rarely successful, and the payoff is nearly always disappointing, when they fail to invest in a formal programme of organisational change management.
According to Gartner, around half of all change initiatives fail, with only 34% being a clear success. The pathway to success says Prosci, is change management. In their ‘An Introduction to Change Management’ paper, they report that initiatives with excellent change management are six times more likely to meet your objectives than those with poor change management.
It doesn’t stop there, though. In the same paper, Prosci advise that with good change management, you are also:
- 5x more likely to stay on or ahead of schedule
- 2x more likely to stay on or under budget
These sorts of stats make change management a must-have, not a nice-to-have initiative.
Is it a matter of public vs hybrid vs private cloud?
No, it’s not. Any one of these options – and not to forget on-premises infrastructure – will support your workloads.
The most crucial consideration is meeting your business goals of achieving greater business value and insights while reducing costs by aligning the right application workload with the right combinations of cloud and traditional infrastructure.
And those workload considerations should lead the way, rather than a blind insistence on cloud-only at all costs.