Esri’s Mark Bramer introduced the Esri ArcGIS spatio-temporal big geo-data store. ArcGIS 10.3 introduced a ‘managed’ relational database designed to capture inter alia, time series data. In 10.4, this has been upgraded to handle high volume and velocity, streaming real time data at ‘tens of thousands of writes per second.’ The system is connectable to a variety of real time data streams including Hadoop/Kafka, MQTT, MongoDB and even Twitter! Tests on the system have demonstrated single node write through bandwidth of over 100k/sec as compared to 0.2k/sec for ArcGIS relational.
In another presentation, DigitalGlobe’s Harsh Govind introduced a new ArcMap to Cloud connector for ‘geo-big-data.’
Infosys’ Sethupathi Arumugam described a novel approach to capturing Witsml drilling data to ArcSDE for real-time monitoring and risk assessment. The idea is to expose generally underutilized spatial information in the Witsml data stream allowing for its use in geosteering and drilling reporting. Infosys has extended Witsml with a ‘risk object,’ an XML description of e.g. fluid losses over a given interval. The risk object adds information on severity, likelihood and possible mitigation measures. In well planning, multiple wells and risk objects can be assessed simultaneously.
The geo-big data theme was also a highlight of the 2016 Esri plenary UC in San Diego with a different slant on spatio-temporal analytics. Esri is working on GeoAnalytics, a new server capability that allows spatial analysis to be run as distributed computations across a cluster of machines. The demo covered a real time data feed of billions of financial transactions. GeoAnalytics was used to aggregate and investigate this large and complex data set to spot fraud. The system spotted multiple near simultaneous fraudulent transactions in different localities, all just below the legal reporting threshold, a likely indication of a money laundering operation. Watch the GeoAnalytics video and download the US PUG presentations.
© Oil IT Journal - all rights reserved.