January 2021 Summaries

Charles Boicey, Chief Innovation Officer at Clearsense, emphasizes the importance of integrating comprehensive patient data, including external sources, to enhance healthcare outcomes. He advocates for a modern, service-oriented architecture using tools like Kong Gateway and Kuma service mesh to manage data efficiently across hybrid environments comprising private, public cloud, and on-premises systems. Boicey outlines a sophisticated data pipeline that includes hyper-converged and composable infrastructure, Hadoop for data landing, and API-centric applications to process healthcare data effectively. He highlights the necessity of applying AI for insights, robust data governance, and schema application to transform raw data for endpoint integration. Through case studies like Meharry Medical College and UCI Future Health, the article demonstrates the potential of data-driven healthcare in addressing disparities and improving individual health management by leveraging personal health data, environmental factors, and streaming analytics, ultimately aiming for personalized patient interventions.

Kuma 1.0.6 introduces significant enhancements, including the first chart visualizations in its GUI and deep status checks for data plane proxies, services, and Kuma's own services, which help pinpoint issues when they arise. The release also adds five new load balancers—least request, ring hash, consistent hash, random, and maglev—to complement the existing round-robin option, allowing for differentiated load balancing rules across traffic paths. Support for Envoy xDS V2 is deprecated in favor of xDS V3, and performance is improved for data storage in Postgres under Universal mode. Users are encouraged to upgrade after consulting the Upgrade Guide and to engage with the community through calls and online channels to further explore and contribute to the CNCF Sandbox project.

In the current era of rapid technological innovation, significant advancements are anticipated, such as humans landing on Mars and the widespread adoption of virtual reality, with the pace of change accelerated by the COVID-19 pandemic. This rapid evolution in technology, originally expected to unfold over a decade, is now reshaping industries much sooner, with digital adoption and customer interaction trends leaping forward by three years. Companies that have embraced innovative processes are reaping benefits, while others face pressure to adapt quickly or risk obsolescence. A 2021 survey of 400 senior technology decision-makers highlighted the importance of distributed application architectures, with 87% endorsing microservices-based applications as key to future success. However, a disconnect exists between executives' confidence and the readiness perceived by architects and developers, suggesting a need for increased investment in scalable technologies to meet rising digital expectations. The urgency for innovation has intensified, with a growing number of technology leaders recognizing that failure to innovate could render organizations obsolete within three years, underscoring the necessity for businesses to continuously adapt and innovate to survive and thrive in an increasingly digital world.

Rate limiting is a critical mechanism for managing API access by controlling the frequency of user requests to prevent system overuse, whether accidental or malicious. This practice is essential for maintaining service quality, especially for public APIs that handle computationally-intensive tasks or sensitive data, as it prevents resource starvation, manages costs, enforces user quotas, controls data flow, and acts as a security measure against various attacks like DoS and scraping. Several algorithms, including Leaky Bucket, Fixed Window, Sliding Log, and Sliding Window, offer different approaches to implementing rate limiting, each with its advantages and drawbacks in terms of handling request bursts, boundary conditions, and scalability across distributed systems. Implementing rate limiting in a distributed environment requires synchronization policies to prevent users from exceeding global limits when requests are sent to multiple nodes, often using a centralized data store to manage counters, although this can introduce latency and race conditions. Tools like Kong API Gateway facilitate the setup of scalable rate-limited services by providing configurable plugins that support different algorithms and data synchronization options, offering flexibility in managing API traffic and ensuring reliable performance even in large, distributed systems.

The Terraform API gateway module facilitates the implementation of Kong using infrastructure as code (IaC) while adhering to DevOps best practices, leveraging Terraform's declarative approach. This module, maintained by Kong, serves as a reference platform for setting up Kong Gateway and Kong Enterprise quickly, particularly for demos and proof of concept environments, but it can also be extended for production needs. The process involves defining a Terraform provider for AWS, configuring the necessary prerequisites like VPC subnets and certificates, and using Terraform commands to provision and manage resources while maintaining statefulness for updates. Security practices are emphasized, including the use of AWS S3 for encrypted state storage and the AWS Systems Manager Parameter Store for sensitive data management. Additional configurations are conducted using EC2, Cloud-init, and IAM roles to ensure secure and efficient deployment and management of Kong instances, including features like auto-scaling, health checks, and API configuration.

Transitioning to microservices offers teams building large applications the ability to innovate and deploy rapidly while enhancing security through zero-trust architecture. This shift from monolithic codebases allows for the compartmentalization of resources into individual services, each exposing an API over the network, which can be secured by implementing a zero-trust model. Zero-trust security requires assigning identities to services and setting permissions, which can be efficiently managed using service meshes like Kuma. Kuma, an open-source service mesh, leverages the Envoy proxy to support zero-trust security across multi-cloud and multi-cluster environments, simplifying the implementation of security policies such as mutual TLS (mTLS) and Traffic Permissions. By using Kuma, organizations can automate TLS certificate issuance and rotation, allowing application teams to focus on building services while infrastructure architects maintain control over connectivity without needing to develop bespoke network management solutions. The adoption of service mesh patterns enables reliable and portable security implementations, ultimately increasing the efficiency of application development and deployment.

In the realm of microservices, efficient cloud service connectivity is vital for developing user-focused products, but achieving this can be time-consuming and complex, often leading to technical debt and security issues. Developers face the choice of writing extensive code for smart clients to manage service connectivity or relying on architecture teams, which can be unreliable. A third, more effective option is to use an open-source service mesh like Kuma, which simplifies connectivity by providing a universal control plane that manages service interactions across various infrastructures. Built on Envoy, Kuma enables seamless and secure service communication, reducing the need for manual configurations and improving developer productivity. It operates at multiple network layers to ensure that all service requests are reliably handled, and it features a RESTful API, GUI, and CLI for easy management. For enterprises seeking enhanced features and support, Kong Mesh extends Kuma's capabilities, offering a robust solution for tackling connectivity challenges at scale. Implementing a service mesh like Kuma can streamline connectivity management, ultimately allowing developers to focus on their core task of building applications.

Kuma 1.0.5, the first release of 2021, introduces performance enhancements and critical fixes, particularly for Secret resources in Kubernetes and TCP traffic logging policies, alongside adding fault injection support for HTTP/2 and gRPC. Users are encouraged to upgrade to this version and consult the full changelog for a comprehensive list of features and updates. As a CNCF Sandbox project, Kuma fosters a neutral, open, and inclusive community, inviting users to participate in discussions and contribute to the project roadmap via official Slack channels and community calls, with the next scheduled for January 20th at 8:30 am PT. Users are advised to review the Upgrade Guide before installing Kuma 1.0.5, and they are encouraged to engage with Kuma on Twitter and GitHub.

Ivan Rylach, a staff software engineer at Checkr, discusses the transition to a services-oriented architecture and the adoption of a declarative configuration for API management to enhance scalability and efficiency. Checkr, a leading company in the background check industry, initially used imperative management but faced challenges including inconsistent API gateway configurations and manual setup errors. To address these issues, Checkr integrated Kong Gateway and switched to a declarative approach, leveraging tools like Kubernetes and Helm to streamline deployments across regions and environments. This shift enabled Checkr to implement a hybrid API architecture, maintain consistent configurations, and automate deployments, while also accommodating consumer management with tools like Kafka. The system also allowed for multi-region availability and ensured eventual consistency in API management, supporting the company's growth by meeting scalability and operational efficiency requirements.

In the wake of the COVID-19 pandemic, 2021 saw organizations prioritizing digital transformation, leading to significant trends in technology and IT management. IT architects gained prominence in decision-making due to their expertise in digital transformation, resulting in more budgetary power for technology purchases. Service mesh technology gained traction for simplifying IT infrastructure, while zero-trust security models became essential as organizations moved to distributed architectures. WebAssembly emerged as a key technology for its portability, overshadowing serverless computing, which was expected to regain importance by 2030. GitOps, combining Git workflow with CI/CD benefits, became mainstream as infrastructure-as-code practices grew, emphasizing Git as the source of truth for organizational systems. These shifts highlighted a broader trend towards more secure, efficient, and architecturally sound IT environments.