It’s well known that if you go noSQL, you lose the relational set operations of a traditional RDBMS. It then falls to the application layer to perform the set joins as needed, or to create a hybrid noSQL/RDBMS environment, but as anyone in big data can attest, it’s just not scalable. It’s possibly the largest problem facing big data. 

Or, at least, it was before Infinitis' UNITY.  

We've solved the Big Data Equation


Unity works between your data source and destination in a seamless, easy-to-use way. Most implementations require very little effort to reap immediate benefits at scale. 


On-Site or SaaS

Leverage your existing noSQL solution, or spin up a custom Unity cluster to keep the data separated. You're in full control. Or push your data to our cloud service and let us do all the work.


Patent Pending

Unity's patent pending technology solves the relational problem allowing complex, set-based operations to occur on unstructured data sets in near-real-time. Scale like noSQL, analyze like SQL.

Example Use Case: 

Real-Time Threat Detection


Three different threat feeds provide known CnC, Ransomware, and other Blacklisted IPs. How can we detect these bad IPs in syslog and netflow data within one second?

Raw Operations Required (problem space)


(383 trillion) operations per second 

are required to search every field in every document against every other field in every other document. 

What about an Index? Like Elasticsearch?

194074 searches per second

are required to find the bad IPs. While the index significantly reduces the problem, it's still untenable.


1406 operations per second 

and real-time threat detection, within milliseconds.


the only solution to the Big Data Equation

Wait! What about scaling?

Increasing documents:

Document Multiplier
 1X 2X5X
Raw Operations

Increasing Data Lakes:

Data Lakes 2 510
Raw Operations

Assumptions are 1,000,000 netflow documents per day with 96 fields, 125,000 syslog documents with 173 fields, and three threat feeds of approximately 33,000 documents per day with an average of 33 fields per document.  Additional data lakes are estimated as 250,000 documents per day with 25 fields per document. Operation and search numbers represent maximum operations at the end of the day. The number of operations and searches would likely fluctuate throughout the day in actual practice. Operation and search numbers are averaged across the day assuming a uniform distribution of streaming data input.