Breaking Barriers: Using Open Source to Drive Big Data Accessibility

Big Data has become the number one roadblock to organizational insight. Mobile devices, IoT, social media and hundreds of other sources are driving up data volumes 40% annually. If Big Data analytics are to become faster and more efficient, organizations need new solutions. Enter GPU-accelerated analytics. Delivering speed increases of up to 1,000x over CPUs, GPUs are enabling analysts and data scientists to query and visualize massive, multibillion row datasets in milliseconds—far beyond the limits of mainstream analytics tools. GPU-based analytics platforms eliminate the need for laborious data manipulation tasks such as pre-aggregation, downsampling, and pre-indexing.
After this presentation, open source users will be able to implement a solution to assess and interrogate Big Data at speed and scale, without locking into a proprietary approach.

Transcript

1. Breaking Barriers: Using Open Source to Drive Big Data Accessibility Open Source Leadership Summit | Half Moon Bay | March 14, 2019 slides: https://speakerdeck.com/omnisci
2. © OmniSci 2019 Todd Mostak Founder and CEO, OmniSci @ToddMostak todd.mostak@omnisci.com /in/toddmostak/ /tmostak
3. © OmniSci 2019 Barriers to Big Data Accessibility/Democratization • Current platforms are cumbersome (require lots of provisioning work, and often require advanced programming skills) • Non-interactive at scale • Expensive (huge hardware footprint!) and often proprietary
4. © OmniSci 2019 GPUs + OSS = MOAR DATA DEMOCRACY! • GPUs allow users to interactively query, visualize and power data science workflows across billions of records with minimal hardware footprint • OmniSci is heavily optimized to leverage the full parallelism of both CPUs and GPUs • But what about the proprietary part? • Our mission of democratizing big data analytics for EVERYONE impelled us to open source the core of our platform and make it available to ALL
5. © MapD 2018 OmniSci Leverages GPU Parallelism 5 High Memory Bandwidth Native Rendering Pipeline Supercomputer Processing
6. DEMO
7. © OmniSci 2019 7 * open source for single node github.com/omnisci/mapd-core
8. © OmniSci 2019 © OmniSci 2019 • omnisci.com/demos Play with our live demos for yourself • omnisci.com/platform/downloads Download the open source OmniSci Core or a 30-day trial of the full Enterprise stack • omnisci.cloud Get an OmniSci Enterprise edition instance in 60 seconds • community.omnisci.com Ask questions and share your experiences Next Steps
9. © OmniSci 2019 Todd Mostak Founder and CEO, OmniSci @ToddMostak todd.mostak@omnisci.com /in/toddmostak/ /tmostak