At the heart of Guavus SQLstream is a rich and flexible set of operators to efficiently process and analyze incoming data streams.

• Operations include: integration, filtering, parsing, enriching, transformation, windowing, correlation, time-based and geospatial queries
• Auto-optimized time-windows operators: Guavus SQLstream implements declarative windowing meaning that windows describe the scope of the tuples taking part in windowed operations, rather than CEP-approaches that normally view windows as data structures.
• Deterministic processing: Guavus SQLstream is a declarative platform based on industry standard SQL data processing semantics – the language of business data – to assure existing business rules are supported with ease and to eliminate the risks in procedural language platforms.
• Proper event sequencing: Guavus SQLstream has native, continuous incremental, and deterministic support for missed events and lack of events while preserving order of events and enabling action on delayed events, useful for any application where record-by-record correctness for monitoring is critical such as network monitoring, security, and financial.
• Seamless telemetry: Guavus SQLstream continuously computes telemetry data that quantifies behavior and performance. Enabling businesses to leverage existing technology investments, this telemetry is available as a table or a stream for easy integration with any third-party monitoring tool.

SQLstream’s approach to streaming:

• minimizes the time developers spend writing low-level code.
• offers full support for Java functions. SQL UDXes (User Defined Transform) enable Java development of streaming operations, and inclusion of third party libraries and even entire systems in a streaming SQL query.
• reduces system performance latency in the millisecond range.
• offers the fastest time-to-value. Streaming applications take minutes to build and deploy
• ensures continuous run. Sources and destinations can be added or removed and applications can be build, deployed, or updated while data is in motion, without interrupting execution. 

Guavus SQLstream can process over 1 Million 16-byte records/sec per CPU core while scaling linearly to over 16 cores on commodity hardware.  

• hardware optimization: 3x-500x less. 
• minimized losses through early detection of failures and external attacks. 
• implementations use existing database and data warehouse investments 
• development takes minutes: fastest time-to-market
• rich feature set (for stream ingestion/egress, processing, querying, and automated actions) covers the entire stream processing data flow; no additional technologies needed.

StreamLab opens stream processing to BIs and data scientists, who can interactively explore machine data streams and building real-time streaming apps without having to write any code.  Features:

• An intelligent suggestions engine supports users in building complete apps from ingestion to analysis and into action, by simply accepting suggestions made by StreamLab.  
• All of the execution takes place on s-Server, at massive scale.
• With StreamLab, everyone (business analysts, data scientists, or Java, SQL, Python, or Scala developers) can efficiently move from concept to real working application in minutes, as deployment is immediate.

Guavus SQLstream works with any data source, format or destination, concurrently, continuously, and in real-time.

By redundantly computing the results of all stream operations on multiple commodity nodes, Guavus SQLstream achieves arbitrary “x9”s of availability in real-world apps without compromising performance or resorting to exotic and costly hardware.