Technology Illustration

IT'S ALL ABOUT THE TECHNOLOGY THAT DRIVES YOUR DATA

No coding, no clusters, and minimum IT.
SiSense is engineered to make your data
preparation and visualization easy.

start free trial

The Secret Sauce: In-Chip Technology

Bottom Line: Faster & No Size Limitations

Processing data with SiSense’s In-Chip technology is 100 times faster than processing it in-memory. SiSense was designed from scratch with a scalable, memory-optimized columnar database that can comfortably handle terabytes of data and dozens of concurrent queries. These features give it extraordinary performance and also makes it easy for non-techies to join data from multiple sources, build interactive dashboards, and share it all in one click.
SiSense compared to the alternatives

Bottom Line: Faster & No Size Limitations

Using the Entire Memory Hierarchy, Even the Chip

In today’s modern chip sets there is a substantial amount of memory. In-Chip technology maximizes the resources that already exist in a commodity computer, by determining how to best use a machine’s capacity to store, compress and access more data faster.

While other solutions only rely on disk and RAM (in-memory), SiSense takes it to the next level by also utilizing the memory available in the CPU. Since memory found in the CPU is faster than when in RAM, SiSense is able to move data 50-100 times faster than in-memory solutions, making us the next generation of BI.

Using the Entire Memory Hierarchy, Even the Chip

A True Columnar Data Store & Query Kernel

Our technology is developed as a columnar database which slices and stores information as columns rather than as rows (relational database). This capability can pull a column on the disk without pulling the entire table, which is extremely efficient for analytics as most people run queries using only a few columns of data, no matter how big the data set.

For example, if you’re joining a dozen tables with 500 columns, but for a specific analysis you only need 3 columns, sales by customer and by product, SiSense loads only a fraction of the data. Since only the information you need is loaded, columnar operations run 50-100 times faster than with a relational database – making it the perfect technology for BI projects.

In addition, the columnar database is optimized for compression allowing SiSense to minimize the amount of data that moves across the memory. The result: SiSense gives an additional performance boost to run queries at record-breaking speed.

A True Columnar Data Store & Query Kernel

Being Smart about the CPU Cache

SiSense works below the operating system to take full advantage of the fastest available memory: the CPU cache.  With SiSense, calculations are performed inside the CPU, without having to copy data from RAM to CPU. Also, our approach is to decompress data in CPU cache, saving considerable memory bandwidth.

By applying a cache-aware decompression – focusing less on the size of the data and more on the data feed – there’s more RAM available for other resources. SiSense gets its speed because we optimize the movement of data from disk, through memory, into the various layers of CPU cache.

Being Smart about the CPU Cache

Single Instructions, Multiple Operations (SIMD)

When processing queries or running analytical calculations, SiSense leverages Single Instruction Multiple Data (SIMD) instructions, also called vector instructions, from modern chip sets by applying vector algebra to the data. Taking full advantage of these In-Chip SIMD instructions enables columns of data to be acted upon in a single instruction.

The CPU can then process data much faster because its processing many data values in parallel. By using vectorization we are able to calculate multiple operations simultaneously, and give an additional performance boost to queries.

Single Instructions, Multiple Operations (SIMD)

Crowd Accelerated BI: More Queries, Faster

SiSense has revolutionized the ability to scale users and queries by creating and reusing query blocks with a feature called Crowd Accelerated BI. Traditional BI tools don’t gain performance from caching unless identical queries are being requested each time. SiSense instead breaks each query into blocks and caches them individually, collecting the blocks in a “recycle bin”. When a new question is asked, query blocks can be quickly reused by simply pulling them from the recycle bin.

In this way, unlike other tools, the more queries running, the faster the response time. And the more people use SiSense to ask questions, the more efficient SiSense works. Built with a scalable, memory-optimized columnar database, SiSense can comfortably handle thousands of concurrent queries.

Crowd Accelerated BI: More Queries, Faster

Designed for Commodity Hardware

Our high-performance analytics database uses disk, RAM and CPU so efficiently, it can crunch terabytes of data in seconds on something as small as a laptop. SiSense won an award at Strata conference for crunching 10TB of data on a less than a 10K computer all in 10 seconds.

By taking advantage of all CPU technologies, SiSense is able to maximize hardware resources and ensure you are not limited by data size. Even in complex scenarios involving multiple users, big data sets and complex queries, any modern 64-bit chipset is enough to handle it all.

Designed for Commodity Hardware