Conducting your own cloud orchestra: Some tips for effective multi-cloud orchestration
Most organizations feel that they’re not yet achieving the full value of their cloud investments1. Of special concern is managing the multi-cloud environment that more than 75% of companies report they are operating in today2—some combination of a private cloud and one or more hyperscale public cloud providers.
One big reason for falling short of ROI expectations is that, as so much attention is paid to the migration itself, less attention goes to actually orchestrating the new cloud environment—managing and operating all the different components after migration and getting them to work in sync. However, here’s a big secret: Effective orchestration is one of the major areas where you can increase your cloud ROI.
Figure 1 – Operations and Governance Activities Addressed by Orchestration
The enablers of orchestration
Orchestration involves the virtual management of multi-cloud workload operations—things like infrastructure deployment, network, load balancing and patching. Then there are critical governance activities such as capacity management, identity and access management, financial operations and compliance monitoring. (For more, see Figure 1.) The mechanism for management of these functions across multiple clouds is a cloud orchestrator. This is a power set of capabilities that automates the fulfillment of service catalog requests, all accessible in a self-service manner and integrated with your ITSM front-end.
Enterprises face a number of challenges in creating effective orchestration capabilities:
- Uninspired expectations for value: Many implementations don’t have very inspiring expectations about the measurable value they want to achieve, so only a fraction of potential capabilities are being executed. We think companies should look well beyond the cost play and instead look to introduce game-changing, transformative benefits. These benefits must be envisioned upfront and then constantly assessed throughout the journey.
- Insufficient training: Each cloud platform has its own orchestration capabilities, leading to a complex operating environment where multiple tools need to be learned by an operations team. These orchestration tools vary in terms of the skills required, capability maturity and operator experience, requiring additional investment in training and resources.
- Complex technology integration: Many technologies need to be integrated to enable comprehensive multi-cloud capabilities. For example, in our experience, more than 70 different systems have to be integrated in most implementations. To communicate with the entire spectrum of systems, companies need to incorporate multiple sub-nets, firewall configurations, access management policies and other configurations.
- Inadequate focus on ways of working: When implementing new technology solutions, it’s important that companies pay attention to helping their people deal with the changes to processes and their ways of working. If organizational change management is neglected, people will naturally fall back to what they know best, meaning they’ll continue to work based on the previous, persistent legacy operating model. Implementing orchestration requires extensive knowledge of infrastructure on top of strong programming skills and use of enabling technologies. This is where Agile/DevOps approaches and adult learning techniques such as pair programming become so important.
- Budget challenges: Orchestration programs reduce IT costs while improving speed, resiliency and operator experience. However, because it won’t look like they are adding new functions for the business, orchestration investments may not look “glamorous,” so getting and keeping budgets might be difficult.
To overcome these challenges, companies must create a common, multi-cloud orchestration plane – add a simple description of what this is – and establish best practices in cloud operations. That’s a significant undertaking. We recommend that CIOs take the following actions to ensure a successful orchestration program:
- Determine the primary goals you want to achieve: These are usually financial efficiency, speed/agility, and/or business transformation, or some combination thereof.
- Set your priorities: Align your program to your priorities (e.g., reduce costs, increase speed, improve resiliency, increase engagement). The key is to be clear and consistent about these priorities and to be diligent in measuring impact along the way.
- Baseline your current operations: Take inventory of what IT services you currently provide to your user community and how frequently each is consumed.
- Assess talent level: Inventory your current workforce skills, determine what skill gaps exist, and develop options for addressing those gaps through training, hiring, and outsourcing.
- Take an Agile approach: Rather than doing everything at once upfront and risking a long program during which some approaches or technologies may fall out of date, start small and then improve and expand incrementally.
- Leverage adult learning techniques: Focus on building skills and on supporting your people as they learn new ways of working. Build these change management activities into the program by leveraging learn-by-doing techniques to build skills in your teams.
- Focus on one automation priority to start: Based on your priority goals, select one service that struggles to be effective today, then establish an initiative to help that service operate in a seamless, automated manner.
- Capture the value: Measure the productivity lift from that initiative and build a backlog of similar endeavors.
1Sky High Hopes: Navigating the barriers to maximizing cloud value
2Flexera 2021 State of the Cloud Report, Jan. 2021