Analyse how Spark might be usefully employed against Gaffer graphs.

[n3101]

There is a small set of Java Spark operations available in Gaffer, which do not appear to be used and perhaps are unusable.

Spark is a prominent big data technology, so it seems anomalous that Gaffer does not support it. Form an architectural view of how it might be used against Gaffer graphs and what benefit this might bring to customers. Outline the new features that would need to be developed to implement this.