Simplify Cloud Deployment Through Teamwork, Strategy

Cloud simplification is the holy grail that CIOs and IT leaders are after. But the process of simplifying often unwittingly introduces even more complications.

It’s a common story: Organizations move from on-premises data centers to the cloud to achieve simplicity, and they often do -- for a time. They grow and innovate. But as soon as they reach scale, cloud complexity sets in.

Increasingly complicated cloud architectures mean higher costs. Suddenly, organizations take notice of what they see as unmet promises from a TCO perspective.

Cloud deployments have become more complex in recent years thanks to the number of service providers and platforms in the stack as well as cross-platform integration and data interdependencies.

Distribution of Workloads = Cloud Complexity

The distribution of workloads across multiple geographies and cloud regions, which can improve end-user latency and reduce risk of outages, also adds a layer of complexity to cloud infrastructure.

Grace Liu, Seagate’s senior vice president of information technology, points out that deployments are difficult to orchestrate and manage consistently because the infrastructure tools native to each cloud platform are typically designed to operate within the confines of the specific platform.

“Business leaders must work with IT leaders and data managers to create a data strategy that works,” she says. “Data is business currency.”

Keeping that in mind, the goal is to deploy a strategy and architecture that lets business leaders see their stored data -- all of it -- as if through a single pane of glass. In other words, CIOs should be able to look across multiple cloud ecosystems in a seamless manner.

“When beginning to think about deployments, organizations need to put themselves in the customers’ shoes,” Liu says. “The customers want a sound, high-performing infrastructure that supports their business value. Organizations need to keep that in mind when designing the technological objective.”

Key Cloud Stakeholders

Key enterprise cloud infrastructure stakeholders required to guide your strategy include cloud, network and security architects, application development, DevOps leaders, and day-two operations and support teams.

Rod Stuhlmuller, vice president of strategic customer and analyst relations at Aviatrix, points out these stakeholders have traditionally been siloed, especially in large enterprise organizations, often not working closely with each other. “An enterprise cloud infrastructure project is an opportunity to bring these teams together under a common objective,” he says.

To create a basis for productive collaboration, it’s important to build a demonstration environment where each stakeholder can be shown how a multi-cloud architecture can meet their requirements.

“Take feedback and involve each team in making the adjustments they require,” Stuhlmuller says. “Let them get their hands on the solution, leveraging workshop style sessions. This will get everyone comfortable with your proposed solution.”

Simplification Invites Complexity

Liu suggests that when striving for simplification, IT organizations should recognize that simplification of architectures is complex and can be disruptive.

That means it’s important to identify the most opportune time that works for the whole organization.

“When simplifying, don’t just think about components like network switches or storage,” she says. “If you focus on moving or simplifying one component, your simplification can invite a lot more complexity. Think about simplifying whole infrastructure solutions. Align at the solution- or service-level first.”

Stuhlmuller advises that enterprise cloud teams should educate themselves on how networking is done, not only in their primary cloud, but in all public clouds.

This allows them to develop a multi-cloud network architecture that will keep them from having to re-architect when – inevitably -- the day comes when the business requires support for a second or third public cloud provider.

“Cloud teams supporting enterprise scale businesses discover that building with basic constructs quickly increases the complexity and requires resource intensive manual configuration,” he says.

He adds that they are also likely to have a severe lack of day-two operational visibility and control that they need to be successful, which ultimately results in an architecture that binds them to a single cloud provider.

The more data can be employed in multiple environments, the more dynamic and fluid it can be -- and the higher its business value. The nearer the storage to where data is created, the more agility you have.

“Do what you need to remove impediments that stand in the way of enterprises putting their data to use,” Liu says. “Data can flow freely to where applications need it, and where it can operate with minimal latency and resistance.”

Co-location Solutions

Stuhlmuller explains that there are several different co-location solutions that can help enterprises in some aspect of their journey to cloud, including hosting workloads that require specialized underlying hardware or offering high bandwidth network connectivity to cloud providers.

From his perspective, it’s important to understand and architect all aspects of your enterprise cloud infrastructure. However, he warns that many vendors use the same popular terms to describe different offerings. For example, “multi-cloud networking” often means the vendor’s service connects to multiple clouds, stopping at the cloud edge and passing your traffic on to a native cloud construct.

“It’s different in each cloud and it does not mean that they are delivering consistent cloud networking, security, and operational visibility in and across multiple clouds,” Stuhlmuller says.

Connectivity to the cloud is important, and he points out many co-location solution providers offer network connectivity offerings support by cloud service providers, such as AWS Direct Connect and Azure ExpressRoute. “Be sure to understand what each piece of your architecture delivers and what it does not,” he adds.

What to Read Next:

How to Plan a Pain-Free Cloud Migration

Will Edge Computing Kill the Cloud?

10 Cloud Strategies to Avoid Cost Overruns