MongoDB provides extra convenience, as network node failures are processed automatically by both client drivers and the server. In addition to the DBMS architecture specifics, Yandex’s years of experience using MongoDB were also taken into account when designing the system.
Choice of storage type
It is easy to choose the data storage type for each cluster individually. We provide two options: network and local storage. The first option is more reliable and flexible, while the second one provides faster speeds.
Easy to maintain
Yandex.Cloud takes care of maintaining the databases and updating the DBMS to the latest versions. At the same time, you can easily configure the databases, restore them from backups, view logs, and use charts to monitor key performance indicators.
Replicas in different availability zones
If needed, you can easily enable data replication in your database. Any changes that are made to the database are copied to the replicas, including those in different availability zones. You can choose which type of replication to use: synchronous or asynchronous.
Encryption and isolation
In Yandex Managed Service for MongoDB all DBMS connections and backups of the database contents are encrypted using the TLS protocol and GPG software, respectively. Moreover, the databases of different Yandex.Cloud customers are completely isolated from each other. The databases do not use any shared components, so no one else can access the data you have uploaded.
Fast data processing
Yandex Managed Service for MongoDB uses fast NVMe-enabled solid-state drives. The result is that you can expect high performance even when handling large amounts of data.
Any use cases related to machine learning involve storing and processing large volumes of data. At first, this is raw data. Once it is processed, a training set is generated and then learning results appear, i.e., components of a resulting model. MongoDB has a flexible storage system that simplifies these processes.
Caching tasks are handled more effectively when the database is running without overhead, and automatically clears the collection after a specified period. MongoDB in Yandex.Cloud meets both of these requirements.
You can enable information exchange between application components by using intermediate message queues. Many popular queuing engines allow using MongoDB as a backend for data storage.
MongoDB is suitable for most tasks in a data storage cycle. You can also use this DBMS with quasi-structured and even unstructured data. For example, GridFS lets you store regular files, including large ones, using MongoDB.