How to formulate and implement a cloud management strategy: A guide

How to formulate and implement a cloud management strategy: A guide Dipti Parmar is an experienced marketing and technology consultant, helping startups, ecommerce companies and B2B SaaS brands establish thought leadership in their industry, with innovative strategies and digital transformation initiatives. Dipti is also a columnist for leading business and tech publications such as IDG's, Entrepreneur Mag, Adobe's, and Inc. When she's not drinking her team's blood (figuratively), she is busy telling vampire stories to little girls who like Disney princesses. She is @dipTparmar on Twitter.

“If you have not developed a cloud-first strategy yet, you are likely falling behind your competitors.” This was the long and short of a Gartner survey on the worldwide spend on cloud applications and infrastructure, as affirmed by Elias Khnaser, VP analyst at the firm.

Cloud services are increasingly replacing stand-alone or on-premises data center technology at companies across industries or sectors. Affordable public cloud services from Amazon, Google, and Microsoft have leveled the playing field for SMBs – no serious exec questions the ability and necessity of cloud technology to ensure business continuity.

So, while your company might rely on umpteen SaaS tools and IaaS compute resources for everyday operations, do you have a strategy for cloud adoption and management? If not, read on for why you need one and how to go about creating it.

Why do you need a cloud strategy?

If falling behind the competition isn’t reason enough for you, get this: a staggering 83% of respondents – including IT ops, developers, cloud architects, and business executives – to the Flexera State of the Cloud 2020 Report said their organisations were at an advanced or intermediate maturity level of cloud usage, with heavy use of cloud apps and services. This holds true across the board for companies of all sizes – not just enterprise.

Clearly, any mature technology needs a budget and a plan to keep it running and prevent it from causing disruption to operations. What’s more, Flexera also found that over 90% of all enterprise organisations have deployed multiple private, public, or hybrid cloud architectures as part of their cloud ecosystem. While this ensures easy, on-demand availability of resources and apps, it requires multiple business domains to work together.

You need a business-driven decision framework and cloud management model that meet the unique needs of your use cases. It is critical that the whole organisation (and not just the IT department) understands why certain workloads run in the cloud, and why specific cloud apps and services have been developed and deployed in the way they are.

Some of the biggest benefits of having a cloud management strategy in place are:

Cost savings: Just like clouds in the sky, every cloud in computing is different. With a clear strategy in place, you know exactly what functionality, integrations, and compliance you need. This helps you negotiate pricing with cloud providers, and get the best deals on apps and infrastructure without compromising on standards.

Reduced risks: An ad-hoc cloud deployment leaves you vulnerable to the risk of disasters or downtime due to infrastructure or security issues. A multicloud strategy helps you diversify essential services over different environments. For example, you can run mission-critical apps with customer data on a private cloud while using a public cloud provider for routine, functional applications or storing archives.

No vendor lock-in: Vendor lock-in in cloud systems is sometimes thought of a necessary evil. However, you can plan for a balance between portability and functionality by focusing on your operational requirements without committing to a particular architecture, choosing open source technology where sufficient, and sticking to SDLC best practices.

IT architecture considerations

With the variety of public, private, and hybrid cloud systems available today, it is hard for even rookie cloud architects to get stuck. Historically, feature comparison was important because workload placements and ownership of stack elements are the key decision points. You can’t afford to compromise on features, applications, tools, and database capabilities, because that will affect the flexibility of your workload and its ability to scale down the line.

Today, there are alternative models that sufficiently meet any organisation’s needs without adding significant risk in any scenario – the Cloud Native Computing Foundation (CNCF) has defined clear standards that all major cloud providers and vendors adhere to. It makes sense to keep your applications upgraded to the latest technology so that CIOs can decide on (or switch to) the most suitable cloud environment at any time.

There are three basic considerations – or layers – that make up a multicloud architecture:

  • Foundational resources: the underlying compute, network, security, and storage elements of the hardware
  • Workload management: VMs, containers and workload lifecycle frameworks such as Kubernetes or OpenStack
  • Service consumption: applications and APIs that decouple IaaS, SaaS, and PaaS constructs

For these to work seamlessly and transparently as a unified system, you need to build your own application stack. While Amazon, Google, and Microsoft all have end-to-end offerings in the public cloud space, you need a strategy for optimising mission-critical functions, especially where large-scale transactions are involved. With private cloud systems, you have more control over orchestration and automation of resources for your stack.

Deployment and ongoing management

The map is not the territory. Once deployed, real-world applications and multicloud environments tend to have minds (and lives) of their own! Your cloud management strategy needs to address the following factors if you want to keep everyday operations smooth and consistently meet functional objectives:

Real time capacity increases: Many times, applications face usage spikes when you least expect them. With predictable workloads that run on private clouds, you can plan a “cloud burst” – you simply buy additional compute or storage capacity on a public cloud for a specified time duration. This arrangement works well for performance-intensive applications that handle non-sensitive data and ensures optimum utilisation of multicloud resources.

Testing and development: A common scenario in the enterprise is when companies run production applications on a full-featured cloud but use legacy infrastructure or a less expensive public cloud for testing. Your strategy will tell you whether to use a pubic cloud for multi-region capabilities (such as CDN) or a private cloud with virtualisation and data colocation facilities for large-scale testing.

Unified management: A great cloud strategy is characterised by its ability to abstract and unify disparate components of the IT infrastructure. True multi-cloud management is when you can operate diverse public, private, and hybrid cloud systems as a single, cohesive environment. Most cloud management platforms today allow for ‘single pane of glass’ management, which enables real-time orchestration and automatic provisioning and de-provisioning of compute, network, storage, and security resources while abstracting the underlying application and database stacks.

Disaster recovery: Include metrics such as Recovery Point Objectives (RPO), which define sustainable thresholds for data loss and downtime, in your cloud strategy. If you operate a multi-cloud architecture, you can plan better for redundancy by maintaining updated copies of data or mission-critical applications on a different cloud system from another vendor.

Over to you

Cloud adoption comes with the promise of cost savings, flexibility, and even digital transformation. With an effective cloud-first strategy, you can automate workload management and implement DevOps best practices to accelerate innovation and agility in your organisation.

And yet, putting together a cloud management strategy is in no way simplistic. You need to constantly test application and infrastructure performance, monitor crucial metrics, and leverage cloud observability to meet the ever-increasing operational demands of global business. Good luck!

Interested in hearing industry leaders discuss subjects like this and sharing their experiences and use-cases? Attend the Cyber Security & Cloud Expo World Series with upcoming events in Silicon Valley, London and Amsterdam to learn more.

View Comments
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *