Josh Lane interesting article on lambda architecture in Azure. His article focuses mainly on IoT specifics, but this design requires only a few minor tweaks to accommodate other non-IoT contexts. The concepts behind lambda architecture are straightforward and uncomplicated on their own.
Imagine we have a single logical stream of data against which we’d like to perform both just-in-time aggregation of narrow slices of the stream (“what’s the average value of X appearing in the stream over the last 15 seconds?”), as well as batch-oriented, resource-intensive queries that may require large subsets (or even all) of the data we’ve seen in the stream thus far (“what’s the average value of X across the entire stream history?”).
In his article he dives into 4 core key part of lambda:
- A logical, streaming data source
- A speed processing layer
- A batch processing layer
- A serving or presentation layer
Implementation for each layer is explained in detail. Author also describes why he wanted to use serverless architecture everywhere. You will also get access to schemes and example code accompanying this article. Excellent!
[Read More]