MemSQL USG, the wholly owned subsidiary of MemSQL supporting the U.S. government, provides real-time analytics to a number of federal customers. MemSQL supports geospatial data types, a leading factor in the company’s decision to join USGIF in December 2015. Trajectory spoke with Mike Kilrain, president of MemSQL USG, and Gary Orenstein, CMO of MemSQL, to learn more about the company and the power its technology provides to the federal government.

What is MemSQL USG’s background?

Kilrain: MemSQL started its federal business in 2013 while working with several government agencies. In 2014, In-Q-Tel came on board as a strategic investor, and MemSQL began development of MemSQL real-time geospatial capabilities within our existing database. In 2015, the company established MemSQL USG, a wholly owned subsidiary. Our co-founders originally came from Facebook, where they implemented scalable software on commodity hardware to solve for massive growth of data and users. They subsequently left Facebook to build MemSQL and deliver those same capabilities to enterprises of all sizes.

What are MemSQL USG’s core capabilities?

Kilrain: MemSQL enables analytics on changing data sets. Customers want to know what is happening in the moment. MemSQL provides that information by way of an in-memory, distributed database using native SQL (Structured Query Language), the lingua franca for data processing. This allows our customers to achieve real-time performance results with well known SQL-based approaches and tools. The speed and scalability behind the technology is groundbreaking. MemSQL is deployed as software on-premises, in cloud instances, or even bare metal. MemSQL is one of the select database technologies deployed on Amazon C2S today. Our commercial clients also leverage Amazon Web Services commercial cloud.

Who are some of MemSQL USG’s current customers?

Kilrain: MemSQL USG supports a number of agencies in the federal government, most of which are focused on geospatial analytics use cases. In the commercial space, we work with well-known organizations such as Kellogg, Akamai, Pinterest, Comcast, Boeing, and several leading Wall Street firms.

What differentiates MemSQL USG from its competitors?

Kilrain: There are three key benefits MemSQL provides to customers. First, our technology can replace existing batch architectures with real-time ingest and analytics complete with geospatial functions. Data volumes and sources are growing, which requires the real-time workloads MemSQL can deliver. Second, we provide a compelling total cost of ownership that plays to the government’s move to a secure cloud, leveraging commodity hardware at scale. Lastly, we have set up MemSQL USG to operate as a subsidiary with specialized resources to directly support the needs of our government customers.

What research and development initiatives is your organization pursuing?

Orenstein: We are constantly pursuing database innovations like Hybrid Transaction/Analytical Processing (HTAP). HTAP is really about merging database and data warehouse workloads in a single system. Historically, organizations have separated the transactional part from the analytical part and must shift data from one system to the other, also known as ETL—Extract, Transform, and Load. Whenever you have to move data from one system to another to run the analytics, you lose the real-time advantage. With HTAP, every query is accurate to the last transaction, or click, event, or log entry.

Kilrain: Batch processing is the way enterprises have operated for the past 30 to 40 years. You have to wait for your data to be processed, then moved to a data warehouse, and then to get a response. Some of our customers were waiting hours or even days to get a response. We are helping them eliminate latency with real-time data pipelines. Using in-memory processing for these pipelines speeds streaming data so it can be analyzed instantly. To make these pipelines more effective, we created Streamliner, an integrated Apache Spark solution allowing customers to more easily ingest and process data. We also focus on machine learning and business intelligence tools to address the needs of government customers. For example, it is hard to find enough analysts to process all of the incoming data in real time, so we are working on solutions to address things like tipping and cueing, anticipatory intelligence, and real-time geospatial capabilities beyond what our customers have requested.

Does your company have any recent news or announcements?

Kilrain: We released MemSQL 5 earlier this year. In terms of the evolution and maturity of our company, getting to a 5 release is significant. The latest release delivers 10 times the performance gains and includes enterprise features to support large commercial and government organizations. We also recently finished our Series C round of financing, raising an additional $36 million, bringing our total investment capital to $85 million. Also of note, MemSQL has been identified as a visionary in two Gartner Magic Quadrants: Operational Database Management Systems; and Data Warehouse and Data Management Solutions for Analytics. This recognition across both quadrants puts the company in a class that includes SAP, Oracle, IBM, and Microsoft.

What will MemSQL USG feature in its booth at USGIF’s GEOINT 2016 Symposium?

Kilrain: There are two compelling demonstrations we plan to showcase. The first is PowerStream, which demonstrates a real-time simulation of ingest and analysis of data from 200,000 wind turbines around the globe. Wind turbines have 10 sensors attached to them, achieving up to two million transactions per second. We use a graphical interface to demonstrate real-time analytics on data being streamed into the database. The other demo we plan to showcase is TaxiStats, an application we’ve created with our partner, ZoomData. TaxiStats uses simulated pick-up and drop-off data from taxis in New York City to perform real-time monitoring and historical analysis.

Be sure to visit MemSQL USG at booth #1514 at the GEOINT 2016 Symposium and to hear the company’s lightning talk, titled “Real-time Geospatial Intelligence at Scale,” at GEOINT 2016’s pre-symposium event GEOINT Foreword.

Related


Posted by Bill Allder III

Leave a reply

Your email address will not be published.