IT infrastructures have evolved from supporting applications and workloads on-premises to supporting “everywhere infrastructure” — in the cloud, at the edge, in data centers or some combination. As this transformation continues, infrastructure and operations (I&O) leaders must design for the unknown and ensure they are flexible enough to respond to rapid change.

“In many organizations, IT infrastructure has grown from multiple directions at once, resulting in a complex structure of hardware, software, services and providers,” says David Cappuccio,  Distinguished VP Analyst at Gartner. “It's time for I&O to redesign this structure and rethink how IT delivers value to the business.”

New infrastructure strategies must cater to the reality that while IT may not know the enterprise’s future requirements, it must be able to quickly deploy applications and workloads wherever they are needed.

Know what you don’t (and can’t) know

One of the more difficult challenges for I&O leaders is planning for what the business will require of IT in coming years. Rather than focusing on physical infrastructure requirements, rethink infrastructure delivery by designing for business outcomes. This requires acknowledging several unknowns:

- You don’t know where you’ll need IT. Workload placement strategies are now based on business needs and business value. Workloads could be needed in specific regions to improve customer intimacy, with specific providers to support high availability, or even spread across multiple countries in response to data location regulatory and compliance issues. IT infrastructure strategies must account for location flexibility.

- You don’t know how much you’ll need at any given time. Assume that linear growth patterns for infrastructure are no longer valid, especially as enterprises move toward infrastructure delivery patterns based on evolving business needs. Having the ability to rapidly scale up or down is essential.

- You don’t know for how long you’ll need it. Depending on workload and business requirements, an infrastructure node may be needed for years, months or even just hours. Defining these node points, and the requirements and types of workloads or applications that target them, will be critical in reducing complexity and avoiding a proliferation of single-use solutions that can be challenging to manage.

- You must minimize cost. As workload requirements change, physical requirements at any given site may change. The ability to ramp up or down quickly speaks to an on-demand or consumption-based allocation model. 

- You need to deploy solutions quickly, anywhere they’re needed. I&O teams must deploy solutions at speed to get in front of market opportunities or satisfy business demands. The historical model of designing and deploying best-of-breed solutions becomes an impediment to agile infrastructure. 

- Downtime can be a board-level event. With many client-facing workloads, downtime and risk are key considerations when determining deployment models. In today’s world, irritated clients can vent their frustrations on social media, which can quickly turn a one-to-one customer service issue into a one-to-10,000 media conversation, negatively affecting corporate reputations.

- You need to work with existing staff and skills. Even in organizations that are not facing hiring challenges, it can be costly and time-consuming to bring new people on board. In a dynamic environment in which workloads need to be rapidly developed and deployed, existing staff and skills become the most-critical components. Organizational learning is vital to codify automation knowledge, unlearn old practices and enable practical learning environments to share, develop and evolve skills.

Consider each of these factors as you rethink your infrastructure delivery strategy and the value that I&O brings to the business.

Design workload placement strategies based on business values

After considering these unknowns, use a logical mapping of business values against different infrastructure node types to create a baseline for workload placement strategies.

Determine the benefits and issues with each location or node type. Focus on business impact and business value through collaboration and engagement with business partners, and then apply weight factors to each category. Examples include:

- Risk. What is the risk to the business and/or customer if this workload fails? What are the cascade effects on other workloads?

- Performance. Does this workload require high performance being used in queries, business intelligence, analytics and intense research operations?

- Regulatory. Do data residency and local sovereignty laws dictate where workloads need to reside? Are there audit or compliance issues that might dictate workloads remaining on-premises?

When looking at individual applications or workloads, using these same business drivers can help map infrastructure nodes to workload types. By creating these nodal definitions and defining the underlying core architecture of each, I&O can create high-level guidelines to follow as business needs change.

When new business workloads need to be quickly deployed anywhere, these templates will have already been defined, and, by mapping business needs against node delivery types, I&O can quickly deploy workloads anywhere.

Source: Article originally published on Smarter With Gartner