arrow_back Great Expectations : Observability in asynchronous systems using temporal logic
Considerations for East-West Traffic Security and Analytics for Kubernetes Environment
Submitted by Akshay Mathur (@akshaymathur) on Wednesday, 27 February 2019
Organizations are moving towards microservice-based architectures across public and private cloud, as well as data centres. As traditional tools are not suitable in containerized environment, there is a need to come up with a new-generation load balancing solution with the ability to address the following pain points:
- East-West traffic security policy enforcement
- Visibility and analytics for East-West as well as North-South Traffic
- Advanced application-layer load balancing
- Keeping the configuration of the load balancer always synchronized with dynamic container environment
In this session, we will discuss both the technical challenges in Kubernetes environment, as well as certain business challenges faced by enterprises while moving to Kubernetes.
Following is the detailed list of challenges to be covered in the session:
A. Technical Challenges for organizations deploying applications in Kubernetes:
Although Kubernetes addresses challenges relating to application build and deployment, organizations still face the tough and often time-consuming challenge of reliably delivering runtime security and performance assurance in containerized microservices environments. It is also important to note that this issue will only become more complex as organisations grow in scale and data.
- As Kubernetes internal and external networks are isolated and Kube-proxy has limited capability, advance load balancer is needed as replacement
- Constantly changing Pods (and IP addresses) pose a challenge for configuring any traffic management and security tool that works on IP address
- There is no simple way in Kubernetes to implement access control between microservices
- Visibility (at application layer) into traffic flowing through microservices is absent
- Implementing side-car proxy significantly increases the resource requirement
- The absence of a container-native enterprise grade solution for application delivery
B. Business Challenges faced by enterprises
With the overwhelming amounts of data being handled by organisations, it makes sense to use a centralized policy management tool. There also exists the need for comprehensive metrics, logs and analytics to provide actionable insights for each microservice. The goal of these analytics is to provide end-to-end transaction response time, throughput, request rate, and other real-time and historical metrics. But, in the process of adopting a solution to tackle these problems, the following issues need resolution.
- A lot of customers don’t want to move to public cloud but want to modernize their infrastructure and potentially adopt a multi-cloud strategy.
- various choices of load balancers (traditional hardware, virtual applinces, cloud native, container-native etc.) confuses the customer in the decision-making process
C. Market Education Challenges in the industry
As the wave of digital transformation takes over every aspect of businesses, organizations are turning to containers for their production applications to become agile and scalable in meeting digital demands that they face. In this journey, what cannot be compromised is the security for both modern multi-cloud and traditional infrastructure.
- With the buzz around containers and Kubernetes, every company has started working in Kubernetes but they are not sure of the architecture and migration plan
- With a lot of marketing of Service Mesh by Google, many think that service mesh can be deploy in only one way and get overwhelmed by the resource requirements
- The approach to tackle all these challenges in one go is selecting a tool for applications deployed in Kubernetes environment, that provides teams deploying microservices applications with an easy, automated way to integrate enterprise-grade security and load-balancing with comprehensive application visibility and analytics – all without changing application code or deployment architecture – and also suites to the use case.
Akshay Mathur serves as Senior Product Manager at A10 Networks. His two-decade experience spans in both technical as well as business sides and in various domains, including wi-fi security, social networking, online retail, cloud applications and application delivery. He is a startup enthusiast and have been founding team member of multiple companies. He loves blogging and teaching on technical as well as non-technical topics via meetups and other platforms.