June 2025 Summaries

Ambient mesh is often touted as a more efficient evolution of service mesh technology due to its leaner, simpler design, and reduced resource usage by eliminating sidecars from every pod. It operates by using a single ztunnel daemon for L4 traffic, which efficiently manages mutual TLS and routing, offering significant CPU and memory savings in high-density environments. However, this resource efficiency introduces operational complexities, especially for L7 traffic that requires centralized Waypoint proxies, adding extra hops and necessitating careful planning to avoid bottlenecks. Sidecar-based meshes, while incurring a higher resource footprint, offer greater operational simplicity, predictability, and isolation, making them suitable for environments that prioritize compliance, team autonomy, and detailed observability. They allow independent scaling and telemetry, reducing operational risks. For mission-critical, multi-zone, or regulated environments, sidecar-based meshes remain the preferred choice due to their maturity and enterprise-grade capabilities, while ambient mesh is more suited for lightweight, L4-dominant applications in single-cluster settings. Despite the promise of ambient mesh, it is still in early stages, and even key proponents like Istio suggest caution in adopting it for mission-critical environments.

In enterprise environments, securely managing LLM credentials is crucial, and one effective solution is to store these credentials in secure third-party systems like HashiCorp Vault or an IDP provider, allowing infrastructure to inject them dynamically. This approach prevents credential leaks, ensures strong governance, and provides the flexibility to switch credentials without redeploying agents. An AI gateway secures the communication between agents and LLMs, and the newly introduced feature in Kong Konnect AI Manager enables the configuration of dynamic authentication using third-party systems, enhancing the security and management of LLM credentials.

Kafka offers significant benefits for real-time businesses, but as usage increases, so does its complexity, often leading to higher costs and operational challenges. To address these issues, organizations can implement strategies such as using Virtual Clusters to provide logical isolation without increasing infrastructure, reducing topic sprawl with centralized and filtered access, and replacing custom code with protocol mediation to streamline integrations. Additionally, applying consistent access control at the API layer and centralizing discovery through AsyncAPIs can enhance security and promote resource reuse. Kong Event Gateway facilitates these improvements, enabling businesses to manage Kafka like a platform, maintaining performance while reducing infrastructure and complexity.

Kong Mesh, built on Kuma, continues to enhance its capabilities with the release of version 2.11, offering improved simplicity and production-grade tooling for distributed applications. It supports seamless multi-zone operations across various environments like multiple cloud providers, Kubernetes clusters, and traditional servers while ensuring zero-trust security and efficient management with Konnect Mesh Manager. The latest update introduces Amazon ECS support with automated Route 53 configuration, reducing the need for cluster roles during mesh setup, and transitioning to an Embedded DNS to optimize resource consumption. It also offers experimental incremental configuration propagation to minimize CPU and memory usage, and expands policy support for MeshHTTPRoute, allowing more refined control over HTTP traffic with features like request timeouts, access logging, and retry logic. These enhancements aim to boost reliability, performance, and manageability for organizations deploying mission-critical services.

Enterprise adoption of large language models (LLMs) is rapidly increasing, with predictions that over 80% of enterprises will use generative AI applications by 2026, up from 5% in 2023. This surge highlights LLMs as crucial tools for business development and productivity, prompting organizations to make strategic decisions about which models to adopt, considering factors like data security and the benefits of open-source versus proprietary models. A survey of 550 Kong users, developers, and IT leaders revealed that spending on LLMs is expected to grow, with current investments exceeding $250,000 annually for some. Despite concerns over security and compliance, newer entrants like DeepSeek have gained traction, with significant usage noted among engineers and tech leaders. Google models also saw a substantial rise in adoption in early 2025, overtaking other competitors due to strategic integrations and aggressive pricing. Overall, 82% of respondents feel that generative AI will positively impact their careers and organizational innovation, although a small minority remain skeptical about its benefits.

Kong's Dedicated Cloud Gateways (DCGWs) have expanded to support all three major cloud service providers—AWS, Azure, and GCP—with a 99.95% SLA, becoming the first API management vendor to enable managed gateway deployments across these platforms. This development addresses historical limitations of hybrid deployments by offering cloud-agnostic deployments with CSP-grade SLAs, enhancing both flexibility and scalability. DCGWs allow for faster market entry, easier scalability, and reduced operational overhead while ensuring robust security and compliance, making them attractive for industries with strict data residency laws. Key features include custom plugin support, smart global DNS, and enhanced observability through advanced analytics and active tracing, allowing organizations to focus on innovation rather than maintenance. Additionally, Kong's APIOps tools, such as decK and Terraform, streamline automation and simplify migrations, providing a comprehensive solution for API management in a multi-cloud environment.

Apache Kafka, a leading event streaming platform, is crucial for building real-time data pipelines and event-driven applications but poses significant challenges when scaling across organizations. These challenges include difficulties in sharing Kafka environments between teams, limitations in supporting non-Kafka developers, restricted external access due to security concerns, complexities in managing configurations, and the lack of documentation which hinders reuse and innovation. Kong Event Gateway offers solutions to these challenges by providing features like Virtual Clusters for logical isolation, event APIs for non-Kafka protocols, enterprise-grade security for Kafka streams, and an abstraction layer that simplifies configuration changes. Additionally, it enhances documentation and observability, facilitating easier access and management of Kafka environments. By integrating with cloud services securely, Kong Event Gateway transforms Kafka into a more accessible, secure, and scalable platform, prompting organizations to reconsider their approach to managing Kafka data with an API platform.

SecureID, a European digital identity platform, addressed the challenge of scaling its infrastructure to accommodate third-party developers by integrating Kong Konnect into its Microsoft Azure environment. This integration allowed Power Dynamic Technology, the team behind SecureID, to rapidly deploy a secure and scalable API developer portal that facilitates the creation of next-generation applications for sectors such as finance and government. By leveraging Kong's high-performance, cloud-native, and developer-friendly features, SecureID was able to expose APIs securely, distribute services efficiently across Europe, and seamlessly support increasing demand. The platform's use of Kong Konnect, Microsoft Azure, and Luxand FaceSDK for biometric security has resulted in a robust digital identity system that is now open for third-party development, exemplifying how strategic technology choices can enhance product delivery and foster innovation.

Kong Gateway Operator (KGO) has released new enhancements to streamline the management of Kong Gateway or Kubernetes Ingress, whether utilizing Kong Konnect or Kubernetes. Key improvements include support for AWS Transit Gateways to simplify cloud networking by centralizing routing, better multi-tenant Kubernetes cluster support with gateway scopes, and the addition of NodePort as a new ingress service type to facilitate developer deployments. The update also introduces a feature allowing users to reference existing Konnect control planes with mirror configurations, easing the integration of KGO with existing setups. Furthermore, the ability to limit the Kubernetes namespaces monitored by KGO enhances resource efficiency and security within shared cluster environments. These changes aim to make it easier for teams with different responsibilities to manage control planes and gateway configurations and encourage users to explore KGO's capabilities with the latest release.