Edge Anomaly Detection
In a transportation company, there may be edge devices collecting instrumentation data on vehicles like trains and cars. Each vehicle could contain a specialized computer with hundreds of sensors reporting various metrics such as temperature, pressure, or throughput for the pipes. There are on the order of 100K data points per second per device. Shipping all these over the internet to the cloud is not practical. As such, monitoring those metrics and identifying patterns or anomalies locally is crucial for the safety of the passengers and operators. There may be also be temporarily loss of access to the internet as vehicles move through different regions of spotty connectivity.
ARCHITECTURE
Without Timeplus:
With Timeplus:
WHY TIMEPLUS?
Lightweight
Single binary, written in C++, extremely efficient with a smaller footprint compared to Apache Kafka, Apache Flink, or Apache Spark.
For high volume use cases, usage can be as small as 8CPU cores and 16GB RAM. Since Timeplus also includes the capabilities of ClickHouse, the battery-included Timeplus Enterprise can greatly reduce the hardware requirements to support hundreds of real-time metrics monitoring at 100k EPS.
Lightning-Fast
Purpose-built to be able to process large amounts of data as well as perform complex aggregations extremely fast in order for the detections not to be stale
Time to detection is key in the success of this sort of application and thus the system has to be able to handle large amounts of incoming data in the shortest possible time to ensure that the detection of an anomaly can be acted upon successfully.
Continuous Running Queries
Purpose-built to have these continuous running queries over fast incoming data to detect anomalous changes.
In contrast, doing this in a database would require custom applications to run these queries periodically to detect the changes. This does not scale and has pitfalls around missing data within windows as well as increasing lag in detections.
Easy Integrations
Native integrations to Kafka and ClickHouse, and connectors to integrate with 200 different systems.
While keeping up with incoming data is important, it is equally critical for the detection results to be pushed out to a centralized system for further analysis and reporting.
Discover End-to-End Capabilities
Bare Metal Installation
Deploy Timeplus directly onto physical servers for maximum performance and control. This setup ensures the lowest possible latency and full utilization of hardware resources.
Data Ingestion
Seamlessly ingest and process data from a variety of sources, including Apache Kafka. This versatility allows for real-time analysis of traffic patterns and potential threats from multiple vectors.
SQL-Based Views
Leverage SQL-based views to create dynamic filters and projections, enabling precise extraction of relevant data from raw traffic logs. Easily identify suspicious activities and anomalies in network traffic.
Windowed Queries
Utilize windowed queries to aggregate and analyze data over specific timeframes or fields, allowing for the detection of DDoS attack patterns that evolve over time or target specific assets.
Shuffling and Sharding
Achieve high-performance data processing with configurable parallelization through shuffle and sharding techniques. Large volumes of incoming traffic data can be ingested and analyzed in real time.
Table Aggregations
Perform large-scale, complex aggregations on traffic data tables to detect patterns. Timeplus handles these heavy computations efficiently, ensuring timely insights into traffic anomalies.
Monitoring
Continuously monitor system performance metrics and business-critical data, for a comprehensive view of network health and potential threats. Timeplus also integrates with dashboard tools like Grafana.
SDKs/Drivers
Access both streaming and historical data through dedicated SDKs and drivers, enabling the correlation of past incidents with real-time traffic for a thorough analysis of DDoS threats and vulnerabilities.
Build Your Detection App with Timeplus Enterprise.
Deploy your way with a 30-day free trial.
No credit card required.
Looking for the cloud?
We've got you covered with our fully-managed cloud service. Rest assured with zero ops, enterprise-grade security, and pay-as-you-go pricing.
Try Timeplus Enterprise Cloud, risk free.