Allocating machine resources

ArcGIS GeoEvent Server can be resource-intensive, true if ingesting lots of events per second, performing complex real-time analyses, or both. Since GeoEvent Server has broad capabilities and solves real-time use cases across almost all industries, you need to deploy GeoEvent Server on a machine that meets your specific use cases. Different use cases require different machine resources.

Machine hardware is continually improving, making it difficult to keep up with the latest technologies available. However, below are some general guidelines for thinking about choosing machine hardware that is appropriate for a typical GeoEvent Server deployment.

CPU

In GeoEvent Server, machine CPU is utilized for real-time event processing and analysis as well as for event data throughput. The speed of the machine’s CPU directly correlates to the number of events per second GeoEvent Server can process. Typically, the more events per second you are receiving, the more CPU is required. Specific CPU optimizations can be helpful such as Intel’s AVX2 technology for processing floating-point numbers.

RAM

GeoEvent Server uses machine memory (RAM) to manage event caches when performing real-time analyses as well as for working with geofences. Capacity for storing geofences in GeoEvent Server is directly related to the amount of RAM dedicated to the machine. The RAM available on a GeoEvent Server machine ultimately dictates how many vertices of any geometry type can be loaded. Also, the speed at which geofences can be accessed is directly correlated with the type and speed of the RAM. In general, if GeoEvent Server will be used to perform complex geofence operations, then increased and/or performance-optimized RAM should be selected.

Network

Typically, in a production environment, GeoEvent Server relies on external connections to receive data. Increasing network bandwidth and reducing latency should be considered when the volume of data ingested is expected to be large. Large in this case means event data that has many fields, lots of data in each event, or frequent data.

Local high-performance drive

When deploying GeoEvent Server, choose an optimal location to store the configuration and message queues. It is recommended to use a local storage device optimized for performance. If choosing a file storage device (NAS/SAN) for this purpose, performance is the most important characteristic to consider. Be sure to learn about event queues and the configuration store before you choose a file storage device.

Choose a storage device that performs well while incurring volumes of small, random input/output (I/O). Consider that read and write performance can greatly fluctuate depending on the characteristics of the I/O. This is an important distinction, since the interaction with the event queues in GeoEvent Server follow this pattern. Oftentimes, this means that a device optimized for large, sequential reads and writes (as often occurs with imagery and video) is unsuitable for use with GeoEvent Server. If your implemented file storage mechanism does not handle small, random I/O well, you may experience significantly increased event processing times or event queue overload.