Tachyon is A Reliable Memory Centric Distributed Storage System developed by Berkeley AMPLab.

CoProcessor with Tachyon is the concept of utilizing this fast storage for processing rather than utilising some other computation engines.

Streaming K-means is algorithm with utilize the concept of forgetfulness to change the update rule in K-means algorithm.

Here are the links where you can study more about Streaming k-means

Spark MLLib http://spark.apache.org/docs/latest/mllib-clustering.html#streaming-k-means

Databricks Streaming k-means blog by Jeremy Freeman https://databricks.com/blog/2015/01/28/introducing-streaming-k-means-in-spark-1-2.html

This post is about the algorithms discussed by Ted Dunning in Spark Summit 2015

Algorithms are:

1. Hashing and Sketching
2. Hyper log log = count distinct
3. Count min = count(s)
4. Streaming k-means
5. Quantiles via t-digest

You can see the video from this link.

Data Stream Algorithms are used in the community for mining large scale real time data for analytics.

The mainly used algorithms in Streaming applications are:

1. Bloom Filter
2. HyperLogLog
3. Count-Min Sketch

Here is the link where you can read more about streaming algorithms.