Scenarios

Offline data processing

Offline data processing is the most common use case of KMR. You can upload raw data to KS3 or clustered HDFS file system, and perform batch offline processing jobs through the console or API.

Instant data analysis

Instant query means you can customize the query according to your needs when you use the system. KMR provides SQL-like query such as Hive and SparkSQL. You can use simple and intuitive query to explore and analyze massive data.

Streaming data processing

Streaming data processing is becoming a hot topic of big data, for example, statistics of website traffic and online game player data should be made at different granularities. This poses real-time requirement, and involves more complex statistical requirements such as the aggregation, deduplication and connection. KMR provides Spark Streaming, a streaming data processing framework, to help you cope with real-time data processing needs easily.