Greeks are famous for many things. But the first thing that comes to mind when we hear “Greek” is Greek Architecture and Greek Alphabets (alpha, beta, gamma…).
Greek Architecture is the flowering of geometry.
What happens when we merge Greek alphabets with Greek architecture? We don’t get new buildings — we get state-of-the-art Data Processing Architectures.
Following are the three popular data processing architectures for big and small data on cloud. These cover various scenarios for both batch, real-time, small, and big data.
Lambda Architecture
Based on the concept of Batch Layer, Speed Layer, and Serving Layer. Lambda is designed to handle massive quantities of data by taking advantage of both batch and stream processing methods.
→ Read the dedicated blog on Lambda Architecture
Kappa Architecture
Based on the concept of processing data in the Speed Layer only. Kappa simplifies Lambda by removing the batch layer entirely, treating everything as a stream.
→ Read the dedicated blog on Kappa Architecture
Delta Architecture
Based on the plan of merging the batch and speed layer. Delta architecture processes any new streaming records as delta (incremental) records, and the data lake is no longer an immutable data structure.