LindormDFS (LDFS) is the underlying storage for Lindorm. It is a distributed file storage service designed for massive amounts of unstructured data. LDFS shares storage with multi-model engines. It combines the low cost and high reliability of Object Storage Service (OSS) with the high performance of Elastic Block Storage. LDFS uses technologies such as tiered storage, replica sharing with deduplication, and intelligent data transfer between hot and cold storage. This provides elastic, low-cost, cloud-native big data storage that is 100% compatible with the Hadoop Distributed File System (HDFS) protocol. You can access LDFS directly using an open source HDFS client.
In scenarios such as data analytics, backup and archiving, and data import, Lindorm allows external systems to directly access the underlying files of its multi-model engines. This greatly improves data read and write efficiency. For example, users can generate physical files in the underlying data format directly in an offline computing system. This reduces the impact on online services.
Core advantages
Low cost: As low as CNY 0.12 per GB per month. Prices may vary by region. See the purchase page for exact pricing. Prices may decrease in the future.
File system foundation: Supports high-performance metadata read and write operations and lets you move files in batches by folder.
HDFS protocol compatibility: You can connect using an open source HDFS client, which helps avoid vendor lock-in. LDFS is compatible with open source ecosystems such as Hive, Spark, Impala, Presto, Flink, Yarn, and Kylin, along with cloud ecosystems such as EMR and MaxCompute.
Decoupled storage and computing: You can purchase storage capacity on demand. Computing and storage are decoupled, which prevents resource waste caused by bundled resources.
Mixed hot and cold data: You can store hot and cold data together and set the data temperature at the file or folder level. This lets you combine storage media with different price-performance ratios to reduce costs.