Ramesh Padmanabhaiah

I have worked with Ashok Srirama many times on critical Kubernetes customer projects. So naturally, when he authored a book on Kubernetes for Gen AI solutions, and asked me to review and write a foreword for it, I was humbled and readily agreed. Here is my honest take: When I first reviewed the manuscript, I was struck by how comprehensive yet hands-on it is. It gives you production grade blueprints on diverse areas - observability to cost optimization, from secure scaling to HA/DR patterns. As someone who spends lot of time enabling teams to scale with Kubernetes and Serverless with AWS, I found myself nodding along and learning from the practical insights this book delivers. Kudos Ashok Srirama and Sukirti Gupta 🙌 Purchase link if interested: https://lnkd.in/ej36jieN

901

28 Comments

One aspect of the work we describe in Systems Correctness Practices At AWS (https://lnkd.in/gBSfXebN) is the work Ankush Desai and team have been doing on closing the loop between formal specification and implementation of distributed systems. When using formal methods in distributed systems, how do you make sure what's running in production matches the specification? What about over time, as the implementation is changed and optimized? Drift between the specification and implementation is a real challenge. PObserve is one great solution to this problem: it allows us to take logs from the system running in production, and check that the behavior they describe matches the formal specification of the behavior of the system. This is a great way to catch bugs, and a great way to ensure that the specification is kept up-to-date. This makes the formal specification work even more valuable.

278

6 Comments

Cross-region data replication with IBM MQ for Kubernetes/OpenShift – it’s official!   IBM MQ is trusted to be the nervous system for many of the world’s most critical solutions, used to deliver trillions of messages each day, with IBM MQ going above and beyond others to ensure that each one arrives safely. But some things are outside of even MQ's control, let's consider a regional failure or even a disaster. The content of many messages are so valuable that you must ensure the highest levels of data resiliency to protect them, and that demands out-of-region replication of your message data, just in case...   IBM MQ Advanced’s Native HA: Cross Region Replication (CRR) builds on top of Native HA, IBM MQ's completely built-in Raft-based data replication and recovery technology. CRR extends beyond a single Kubernetes/OpenShift cluster for HA, adding the ability to asynchronously replicate your live messaging data to a second Kubernetes/OpenShift cluster, outside your region. This gives you an alternative location to fail over to in the event of a problem or for disaster recovery.   Yesterday we announced IBM MQ 9.4.2, this brings many new MQ features and enhancements, including the long awaited CRR. Take a look at the announce and be ready for availability in just a few days’ time.   https://lnkd.in/eUkVAyyJ #IBMMQ #OpenShift #DisasterRecovery #Messaging

502

27 Comments

We recently announced the GA of Oracle Database@AWS. Customers can now run Oracle Exadata Database Service and Oracle Autonomous Database on dedicated infrastructure on OCI within the AWS U.S. East (N. Virginia) and U.S. West (Oregon) Regions. In addition, to meet growing customer demand, we plan to make the service available in an additional 20 AWS Regions around the world. InfoWorld has the details: https://lnkd.in/gswnvM9z

108

8 Comments

My run-mcp-servers-with-aws-lambda library now supports Bedrock AgentCore Gateway! Run your existing stdio-based MCP server in a Lambda function, and let the Gateway take care of implementing the MCP HTTP transport protocol https://lnkd.in/g-Q67G3N

379

8 Comments

OpenSearch becomes more cost-effective by supporting Derived source, with storage cost savings of 40+%. Similar to the newly added Star-Tree index, this is very powerful for reducing cost and improving performance of large scale workloads. https://lnkd.in/gbSZEbVJ #OpenSearch

121

Running APIs in Kubernetes? It’s not as simple as dropping your gateway into a cluster. This post breaks down the real challenges of API management in K8s, from autoscaling and observability to security and GitOps, and shares how Kubernetes-native approaches can help you solve them. A must-read if you’re navigating API complexity in modern environments: 🔗 https://lnkd.in/g8eTaavv #APIManagement #Kubernetes #CloudNative #DevOps #WSO2 #APIGateway

17

Scaling Cloud Architectures: Why Event-Driven Producer/Consumer Design Rocks Looking to build cloud systems that scale effortlessly and stay reliable? The event-driven producer/consumer pattern is the key. Here’s how it works: Producers create tasks or messages and put them in a queue (think AWS SQS).  Consumers pull from the queue and process those tasks asynchronously on their own schedule. This means producers and consumers are not tightly coupled — they can scale independently. Need more power to handle traffic? Just add more consumers. The queue smooths out spikes and keeps things running smoothly. Why it matters: Producers and consumers can scale independently, giving you more flexibility.  Failures in one component don’t bring down the whole system.  You only use resources when needed, optimizing cost and performance.  Features like automatic retries, dead-letter queues, and failure isolation (built into AWS managed services) boost reliability. This pattern fits perfectly with microservices and serverless architectures. For example, pairing AWS SQS with compute like Lambda can handle millions of events efficiently and reliably. 💡 How have you used the event-driven producer/consumer pattern to scale your cloud systems? Share your experiences! #CloudArchitecture #EventDriven #AWS  #Microservices 

45

16 Comments