Amazon Timestream

Timestream is a fast, scalable serverless Time-series Database from Amazon built specifically to store IoT and Operational application data. It is capable of processing trillions of events per day up to 1000 times faster with as little as 1/10th of the cost of a typical relational database. Timestream saves time and cost by intelligently managing the lifecycle of the data by keeping recent data in-memory and moving the historical data into a cost-optimized storage tier based on user-defined policies.

Since Timestream is serverless, it scales the performance up and down on-demand and also has built-in time series functions to analyze the data to find underlying patterns and trends in real-time.

Timestream can be easily integrated with Kinesis, IoT Core for Data collection, Amazon Quicksight, Grafana for Visualization and Amazon Sagemaker for running Machine Learning algorithms on the Time series data.

• Data Lifecycle Management: Amazon Timestream simplifies complex data management in order to reduce overhead costs and increase the performance of the cluster. Timestream offers two storage tiers as follows:
  • in-Memory store: When data is first written to Timestream it is stored in memory which identifies duplicate data values, sort the data, and durably store the recent data. Suitable for querying with fast performance.
  • Magnetic store: We can configure lifecycle policies that upon expiration automatically transfer the historical data to a cost-effective magnetic store and we can configure a retention period after which the data is permanently deleted from the Timestream. Suitable for querying Large volume analytical data.
• Serverless with Auto-scaling: Timestream is serverless and therefore we have no need to procure or manage any servers and their maintenance. Customers can concentrate only on the software and business layer logic without any hassles of maintaining and scaling the underlying infrastructure to cater to varying business needs.
• Built for Time-series data: Timeseries data consists of varying metrics with a timestamp that describes the event changing over time. Such a dataset is used to derive insights into the performance and health of the machine or application, detect anomalies and optimize data transfer. Timestream is explicitly built for such use cases and can quickly analyze data using on-the-fly SQL queries, built-in time series functions for smoothening, approximation, and interpolation of data. It also supports advanced aggregation, window functions on complex data types such as Arrays & Rows.
• End-to-End Encryption: Timestream ensures data is always encrypted whether at rest or in transit between other services by utilizing AWS KMS Customer Managed Keys (CMK) for encrypting data in the magnetic store.

• IoT Data Collection & Edge Analytics: Timestream can be used to store and analysis of complex industrial and IoT Telemetry data to streamline equipment management and maintenance. Such time-series data can be queried and analyzed in real-time to detect and predict anomalies and monitor the overall performance and functioning of the system.
• Website User Interaction Analytics: Timestream can also store User clickstream data and monitor the incoming and outgoing web traffic from the application. We can then use aggregation and approximation functions to determine user interests & the Path to Purchase and Shopping cart discard rate, for example in an E-commerce application.

Amazon Timestream is an impeccable candidate when it comes to storing large scale complex time-series data and process trillions of data points every day to derive rich insights using AI & Machine Learning services like Sage Maker to monitor and predict anomalies in advance. When compared to Dynamo DB or other Relational DB offerings from AWS, Timestream is capable of processing the same amount of time-series data at 1/10^th the cost and up to 1000 times faster which makes it lucrative in IoT, Clickstream Analytics, DevOps, Application Performance Monitoring (APM) use cases.