Design principles embraced for building robust software applications were formed during an era when memory was a primary differentiating attribute, and moderate latency was a perfectly accepted norm. In the cloud ecosystem, the demands on performance are ruthless. This podcast takes a critical view on shaking up the design concepts and how do we critically question and reevaluate our fundamentals

Focus is more on data-centric design and why perfect normalization is no more an assured guarantee for higher performance 

For sharing your ideas, feedback, pls reach out to me at [email protected]

Thankfully acknowledging the contents from the below links used as a direct/indirect reference in this podcast

https://rubygarage.org/blog/

Security a primary factor that holds back people from embracing the cloud, it's a shared responsibility between the vendor and application owner. This Podcast covers a research article from Palo alto networks on how a error message acted as feedback mechanism to try different account ids and hack into the system. The current architecture doesn't cap the number of failed attempts nor there is a timing window that completely facilitates the exercise. Further, the insidious attempts are logged into the account of the hacker completely blindsiding the operations admin 

A similar exercise was successful in the past using a similar concept of IAM Role assumption, though it was quickly fixed by AWS team, the base concept between the two exercises remains the same 

For Questions/feedback/suggestions pls mail me at [email protected]

#aws #security #cloud 

As Cloud competition intensifies, it results in deep discounts to capture market share, also each public cloud comes with its own uniqueness such as GCP for ML, Azure for serverless, Azure for MS-Stack, etc. Both these features explain the rising popularity of the Multi-cloud strategy. However, this also presents a unique set of challenges that are discussed in this podcast, these challenges shouldn't be considered in silos rather it dovetails into one another.

1- Specialty vs standardization - Portability challenges 

2- Lack of qualified tool kits

3- Lack of all-round expertise

4- Compliance and Cost - ITAR, GDPR, SOX,HIPAA,

5- Scaling & Tracking 

6- Cost of Reliability - Overhead 

For Questions/feedback/suggestions pls mail me at [email protected]

#cloud #multicloud #Testing #standards #challenges #Compliance 

This week's Podcast covers, how testing is getting transformed and redefined in the Cloud era. Change is inevitable, so what it takes to pursue and overtake these challenges in order to stay relevant. We discuss about 5 factors which are critical from a tester standpoint in this transition from on-prem to the Cloud

1- Architectural Divergence - From Monolithic to Microservices /Serverless

2- Security as a shared responsibility with application spanning over Hybrid cloud and on-prem

3- Lack of Toolkits and Cloud expertise

4- Diminsihing boundaries between App services and Infra - The compelling need to know the cloud basics

5- Evolution of Polyglot - Inhibiting testers to be bonded to a specific Tech stack 

For Questions/feedback/suggestions pls mail me at [email protected]

#microservices #cloud #cloudcomputing #cloudarchitecture #testing 

This podcast deals with "Security In Cloud" and "Security of Cloud" as well as gets into the details of best practices for safeguarding your Server (EC2). Collection of simple best practices (outlined below) will go a long way in safeguarding your server  

1-Timely Patching of Guest OS

2- Avoiding downloads from untrusted sites

3- Avoiding default Security groups 

4- Isolating your machines(even if it needs to be public-facing) in private subnets and connecting through load balancers

5- Avoiding Open to internet SSH

6- Evaluating SSM as an option than SSH

7- Harnessing WAF - Firewall to detect and prevent SQL injection

8- AWS Shield to prevent DDOS (Distributed Denial of Service)

9- AWS Inspector for accessing vulnerabilities 

10- Handling IAM with the right level of least privilege

The above guidelines are not one time but need to be repeated periodically so that you can stay at top of security 

For Questions/feedback/suggestions pls mail me at [email protected]

#aws #ec2 #security #IAM #cloud #cloudcomputing #cloudarchitecture #awscloud #cloudapplications #performanceengineering

In this concluding part of the podcast, "Pitfalls in Serverless," we continue to discuss two major issues in serverless, this is more to deal with Boundary scenarios of Throttling limits

1- Queue of Death (from 5:50) - SQS routing valid messages to Dead queue as Lambda couldn't process it in the throttling stage, this essentially translates to loss of revenue/business. By balancing Concurrency limit, Max receive count and visibility time out we can minimize the impact but cannot completely eliminate it 

Reference:  https://data.solita.fi/lessons-learned-from-combining-sqs-and-lambda-in-a-data-project/?utm_source=thenewstack&utm_medium=website&utm_campaign=platform

2- Million Dollar Problem (From 18:20) - How a startup saved million dollar in AWS bills by cutting short Dynamo DB partitions and elevated its performance which was dropping down, all due to missing data, a common error in IT applications

Reference:  https://segment.com/blog/the-million-dollar-eng-problem/?utm_source=thenewstack&utm_medium=website&utm_campaign=platform

For part -1 of this podcast, you can refer to this link - https://anchor.fm/tech-padasala/episodes/Caution---Serverless-el1n68

For questions or feedback pls feel free to write to me at [email protected]

 #cloud #cloudcomputing #cloudarchitecture #awscloud #cloudapplications #performanceengineering #aws

Off late Serverless is so hyped up and its benefits of Scale, Agility, and Cost-effectiveness is often overstated. It's not a magic wand that can address all design challenges, as with every engineering concept, serverless does come with a set of merits and demerits. This 2 segment podcast analyses some of the pitfalls in serverless computing. Through 4 such real-life scenarios, we bring out the lessons and best practices to be adopted while deciding on serverless

In this segment we cover 

1- cost escalation due to a simple configuration error 

2- Performance dip between Serverless and traditional server model for a consistent heavy payload

References :

https://sourcebox.be/serverless-a-lesson-learned-the-hard-way/?utm_source=thenewstack&utm_medium=website&utm_campaign=platform

https://einaregilsson.com/serverless-15-percent-slower-and-eight-times-more-expensive/?utm_source=thenewstack&utm_medium=website&utm_campaign=platform

Feel free to write to me at [email protected]

In this segment, we cover strategies for 

a) Quick Recap of last segment - Cloud front and web caching strategy 

b) Application Cache - How to decide what to cache, how to monitor and course-correct contents

c) Data Cache - What to cache, and how long to cache, why invalidation is an important strategy, Leveraging Lambda to synch Cache and DB. Options available in the cloud for the cache services 

Reach to me at [email protected] for feedback, clarification, and comments 

Building High performing architecture using Cache, the way we leverage cache will not only optimize the performance but also saves cost and elevates user experience. In this Podcast, we deep dive into the less familiar features of CloudFront and Lambda at edge. We also cover Web caching as an insulating layer to guarantee the performance. Segment-2 of this podcast we will continue with App Caching and Database aspects in detail

As a continuation to my previous podcast on S3 optimization Techniques, we cover the following topics in this podcast

1- Connection Pools

2- Byte Range Fetch

3- Retry Requests and Back off strategy 

4- Leverage Latest AWS SDK

5- Cloudfront and Elastic Cache 

Pls share your feedback/query to [email protected]