Service Mesh

Survey Results: Service Mesh Useful for Security, Observability and Traffic Control

27 Feb 2020 9:22am, by

People familiar with service meshes say the technology is important to organizations’ efforts to improve the security, observability and traffic control of distributed systems over the next twelve months. That is one takeaway from 138 responses to an online poll The New Stack conducted Feb. 13-24. Access to the raw data and tabulated results are in a publicly available workbook.

A third of respondents’ organizations are using service meshes to control communications traffic between microservices in production Kubernetes environments. Another 34% use service mesh technology in a test environment, or are piloting or actively evaluating solutions. Instead of trying to gauge actual adoption, we are using this data to explore how decisions are being made in the near future.

Service meshes have yet to become adopted by “early majority” technology adopters, but 46% of the survey are piloting them or have plans to evaluate or implement them in the next 12 months. Stories of successes and failures in production environments may affect these plans.

Among companies using services meshes for production Kubernetes environments, 60% believe they are extremely essential for near-term improvements controlling application traffic within distributed systems, and another 30% say they are important but not essential. The figures are only slightly lower in terms of technology’s impact on security and observability. Among those kicking the tires with service mesh, 51% say it is essential for security improvements and 43% feel that way regarding observability. It is possible that security and observability are viewed as one of many technologies that address what continues to be two of the biggest problems facing the DevOps and site reliability engineering communities.

Future research is needed to measure the technologies’ impact on enterprise-wide distributed systems.

Only 30% of the evaluation/piloting/using in test cohort say meshes are essential for improvement to their organization’s traffic control, but we expect this figure to rise dramatically as people get their hands dirty with the day-to-day operations of the technology.

The mini-survey concluded by asking for input on the angles The New Stack should take as it continues its coverage of service meshes. Adoption of service meshes in non-Kubernetes environments is the least popular suggested topic, which is partly because the first question in the poll narrowed in on the Kubernetes use case.

People at organizations with production use of service meshes were most likely to be interested in two topics: 1) 62% want to learn more about how to control security policies in a hybrid environment (e.g., multicloud, multiple service meshes, non-container, other service meshes), and 2) 51% want to hear about companies that have investigated open source service mesh options, such as Istio, but end up deciding to develop their own mesh capabilities or bring in proprietary enterprise solutions. Respondents that are using service meshes, but not in production, are most interested in two other topics: 1) preference for pluggable interfaces (bring your own) vs. service meshes that include ingress control and/or API gateway, and 2) the importance of integrating control and data plane components into a single solution.

One respondent wrote asked for “real customer stories, even if they are failing. The hype is annoying.” Another wants more education about the “importance of mesh to devs and QA teams.”

We encourage readers to use this research in future contributions to our website or as a topic for an upcoming podcast.

Results Based on Extent of Adoption


Feature image by John Mounsey from Pixabay.