1) Spark 1.0 was released in 2014 as the first production-ready version containing Spark batch, streaming, Shark, and machine learning libraries. 2) By 2014, most big data processing used higher-level tools like Hive and Pig on structured data rather than the original MapReduce assumption of only unstructured data. 3) Spark evolved to support structured data through the DataFrame API in versions 1.2-1.3, providing a unified way to read from structured sources.