共享服务平台是否过于限制？_AI阅读总结 — 包阅AI

The belief that shared service platforms are too restrictive and inflexible is expected to be the top challenge facing platform teams, according to a report released Tuesday.

The perception has grown alongside increased acceptance of shared service platforms, reported the study by Rafay Systems. While only 28% thought this was a challenge last year, it is the top expected challenge, cited by 44% of respondents, platform teams will face in the next year.

The study also indicated widespread adoption of platform engineering. Seventy-one percent of the participants in the survey said their organization has a platform engineering team, while another 20% have an unofficial group that acts as a central platform team.

Platform teams are expecting significant growth, according to the report. Sixty-nine percent of organizations with platform teams expect that staffing for platform engineering will grow 25% or more over the next year. Almost as many (65%) expect at least a 25% growth in their platform budget over the next year. However, only 47% of platform teams have a budget that is separate from IT.

It is important to realize that platform teams are likely more prevalent among survey participants than among companies worldwide, due to the study’s sample.

The study’s participants were recruited through a third-party panel provider. All of them work in the U.S. for a company that has at least 1,000 employees and has experience with Kubernetes. Fifty-two percent had roles primarily dealing with IT/cloud, 15% focused on platforms, with the remainder holding development, DevOps and site reliability engineering (SRE) roles.

The report did not identify what budgets, teams or job roles will be negatively impacted by platform engineering trends. For example, will cloud and DevOps teams lose staffing as more workers join platform teams?

Platform Team Challenges

Platform teams themselves are facing a variety of challenges, with 43% of respondents noting difficulties with the compatibility and integration with existing tools and workflows. Furthermore, 39% said their organization’s platform team has been challenged by the complexities of provisioning and managing multicloud environments.

Beyond operational concerns, perceptions, and possibly misconceptions, about platform engineering continue to be roadblocks for platform teams.

Forty-one percent of the survey said platform teams are challenged by the fact that many other teams still prefer to build and manage their own services or tools. No wonder 31% believe that platform teams have been challenged by a lack of awareness or understanding of the benefits of a shared services platform.

However, it appears that there is a growing acceptance of the need for platform teams. Looking at possible challenges in the upcoming year, fewer believe they will include a preference to build independent services (35%) or a lack of awareness of shared platforms (26%).

Source: Rafay Systems’ “The Pulse of Enterprise Platform Teams: Cloud, Kubernetes and AI”

What Slows Down Environment Provisioning?

Multicloud and environment provisioning are unique concepts. Multicloud provisioning is about managing resources and user accounts across multiple clouds in order to improve scalability and flexibility. Meanwhile, environment provisioning deals with setting up and managing a single cloud environment to support application deployment.

Half of the survey their organization has a platform team working on an initiative related to multicloud and hybrid cloud environment provisioning and management. And as stated previously, multicloud environments can be complex to manage.

When asked about their developer and automation efforts, 28% said their organization has an initiative to automate environment provisioning and management. That’s less than the 40% building internal developer platforms and 45% focused on providing self-service experiences for developers.

Among the 28% who reported environment provisioning and management initiatives, Rafay found that:

• 82% said it takes at least a month to deploy a cloud application/service into production once the code has been completed and tested for quality assurance. That is much slower than desired, as 80% believe it should take less than a week to go from completed code to deployment.
• When asked what is causing delays in moving completed code into production, 61% mentioned endless cycles of troubleshooting at the infrastructure level, 56% noted the complexity of environment provisioning and management, and 46% the back-and-forth between application and infrastructure teams about infrastructure requirements.
• Lower on the list of delays were waiting for someone to provision or spin up environments (43%) and onerous ticketing systems for provisioning (26%).

Source: Rafay Systems’ “The Pulse of Enterprise Platform Teams: Cloud, Kubernetes and AI”

Self-Service for Developers

Looking forward, 44% of the survey plan to prioritize self-service for developers over the next year. Among those focused on self-service, the report found that:

• 65% want to enable self-service workflows for Kubernetes cluster provisioning.
• 60% want to enable self-service monitoring for applications (60%) and infrastructure (51%).
• 47% are prioritizing namespace as a service (47%) and self-service for environment provisioning (43%).