Why customers choose Oracle Autonomous JSON Database over MongoDB

Is the service optimized for database usage?

Oracle's binary JSON (OSON) format is optimized for efficient database operations. Its offset information allows “jumps” to selected values. Piecewise updates limit transaction log size, and a compression scheme reduces the JSON document’s physical size.

MongoDB's BSON format requires more IO when reading and writing data. In addition, the BSON format is limited to 16 MB, requiring greater segmentation of data into multiple documents, compared to Oracle’s support for 32 MB documents.

Does the service provide full SQL support?

Oracle Autonomous JSON Database combines the best of JSON document stores with SQL and provides modern document APIs. With Oracle Database 23ai, users can run Oracle SQL statements within the MongoDB API.

MongoDB lacks SQL functionality, which would require developers to write many lines of code, increasing development and maintenance costs and creating security vulnerabilities.

Does the service provide comprehensive relational capabilities?

JSON Relational Duality, a feature of Oracle Database 23ai, allows developers to have data in JSON format. Which means that they do not have to choose between the benefits of relational capabilities and JSON capabilities.

With MongoDB, data needs to be duplicated to keep it consistent with complex application code. Running analytics and reports against the data is a challenge in MongoDB.

Does the service support other data models than just JSON?

Autonomous JSON Database offers full multimodel capabilities, including Full Text search over JSON, GeoJSON, XML, Graph as well as a new AI Vector data type with special indexes to support new AI workloads.

MongoDB lacks multimodel database capabilities, requiring developers to stitch multiple separate databases to build modern applications. This results in slower application development and applications that are prone to security vulnerabilities.

Are optimized machine learning capabilities built into the database?

Autonomous JSON Database provides capability to run ML inside the database. A fully integrated AutoML user interface is provided within the JSON autonomous database.

MongoDB lacks built-in machine learning algorithms, requiring separate service for ML. Using multiple services increases management complexity and creates potential security gaps due to data fragmentation.

Does the service support low-code application development?

Autonomous JSON Database is preconfigured with Oracle APEX, a highly productive, no-code/low-code environment.

MongoDB does not have an equivalent of Oracle APEX. To implement applications with low-code tools, developers and business users will need to use third-party tools, learn to use them, and incur additional costs.

Does the service automatically scale compute nodes of the database cluster?

Autonomous JSON Database uses scalable, high-performance compute and storage resources for all types of workloads and can elastically scale the size depending on the workload.

MongoDB developers must decide on cluster type as well as their cloud vendor. Further, changing the cluster type is not easy with MongoDB workloads.

Does the service support unlimited storage for JSON collections?

Autonomous JSON Database does not have a storage limit for JSON collections. Organizations can create enterprise-wide document databases without worrying about running into capacity constraints or having to increase configuration complexity. In addition, storage is provisioned separately from compute, so you can minimize costs for very large Autonomous JSON databases by only using a few vCPUs.

MongoDB storage is dependent on the type, size, and IaaS provider that the cluster is running on. Customers must pay for the additional compute that comes with the storage, even if they don’t typically need it. The maximum MongoDB document store is 4 TB—which comes with 96 vCPUs. However, clusters with 8 vCPUs may be limited to document stores as small as 128 GB. Expanding beyond MongoDB’s 4 TB limit (in fact 4000 GB) requires customers to use sharding across multiple databases, increasing complexity and cost.

Does the service auto-tune itself to deliver optimal performance based on dynamic workloads?

Autonomous JSON Database automatically configures and tunes itself as data and schema change over time. Memory configurations, data formats, indexes, and access structures are automatically optimized to improve performance. No DBA involvement is needed.

MongoDB does not provide any auto-tuning capabilities. Developers or DBAs must manually tune multiple aspects of the database, including creating indexes, searching for storage wait states, locking performance, and cluster overload. Manually tuning the MongoDB database takes time and effort and may result in suboptimal performance as workloads change.

Does the service automatically repair itself?

Autonomous JSON Database continuously monitors hardware and software to identify faults, predict failures, and resolve issues before they impact database operations. Automatic SQL plan management regularly detects and repairs SQL performance regressions.

MongoDB reconfigures itself when faults occur, and it does not offer immediate failover or application transparency. MongoDB typically pauses for 12 seconds during cluster failovers and elections, forcing customers to add retry logic to their applications to handle the delays.

Does the service deliver high performance without sacrificing ACID transaction consistency?

Autonomous JSON Database is built on the proven Oracle Database platform, using Oracle Exadata infrastructure to provide low-latency create, read, update, and delete (CRUD) operations and maintain full ACID transaction consistency for enterprise applications. Autonomous JSON Database eliminates data corruption and data loss when performing a sequence of write operations on multiple documents as a single transaction.

MongoDB is not a transactional database. Its support for multi-document ACID transactions was only introduced in May 2018. MongoDB ACID capabilities come with many limitations that come up short on supporting consistency. For example, it is recommended that MongoDB users not update to more than 1,000 documents in a transaction to limit the possibility that the transaction will be aborted due to timeouts. Developers are forced to work around these limitations by writing their own bulk updates with error checks, commits, and rollback logic. Third-party reviewers also confirmed the limitations of MongoDB’s ACID transactions.

Does the service deliver superior performance?

Autonomous JSON Database’s native JSON binary format, OSON, is highly optimized for fast reads, avoiding linear scans and partial updates while reducing redo/undo log sizes. As a result, Autonomous JSON Database significantly outperforms MongoDB. As the performance benchmark demonstrates, Autonomous JSON Database is up to 3.2X faster than MongoDB for workloads represented by the YCSB benchmark.

MongoDB binary format, BSON, requires linear scans and full document updates, causing it to run slower than Autonomous JSON Database. In addition, documents stored in BSON typically take up more storage than documents stored in Oracle’s OSON format.

Does the service automatically scale itself on a granular level to optimize query performance and concurrent user throughput while limiting costs?

Autonomous JSON Database automatically scales compute resources to optimize query performance and concurrent user throughput requirements. Compute resources are automatically and instantly scaled down when they are no longer needed. Compute and storage scaling take place independently, so increasing storage capacity does not force higher compute costs. Organizations only pay for the minimum level of compute resources they consume.

MongoDB scales clusters to the next sizing tier, but this scaling is neither granular nor instantaneous—and differs based on which IaaS provider a customer selects.

Does the service include a comprehensive way to monitor and manage security?

Autonomous JSON Database comes with Oracle Data Safe at no additional charge. Oracle Data Safe helps customers understand the sensitivity of their data, evaluate risks to data, implement and monitor security controls, assess user security, monitor user activity, mask sensitive data, and address data security compliance requirements.

MongoDB has basic security features, such as authentication, authorization, encryption, network isolation, and database auditing, but lacks the equivalent Oracle Data Safe functionality. Organizations must implement Data Safe-like features using third-party tools to secure their data, increasing risk and operational and administrative costs.

Does the service provide security controls to prevent privileged users from accessing sensitive data?

Autonomous JSON Database comes with Oracle Database Vault, which provides powerful security controls to protect application data against unauthorized access to address privacy and regulatory requirements. Controls can be deployed to block privileged account access to application data and control sensitive operations inside the database.

MongoDB does not offer this capability and currently cannot provide preventive controls to block privileged users and DBAs from accessing sensitive data in the database.

Does the service support extensive government and industry certifications?

Autonomous Database supports a broad suite of compliance attestations and certifications, including HIPPA, ISO/IEC 27001:2013, ISO 27017:2015, ISO 27018: 2014, SOC 1, SOC 2, FISC, NISC, and NCSC guidelines. In addition, Oracle Cloud Infrastructure has been successfully evaluated for a number of operational standards.

In contrast, MongoDB provides a very limited set of certifications, including HIPPA, SOC 2, and ISO/IEC 27001:2013.

Can the service be deployed in customer data centers to meet data sovereignty requirements?

Oracle Autonomous Database supports a wide range of Autonomous JSON Database features. It is available in customers’ data centers on Oracle Exadata Cloud@Customer, allowing customers to gain cloud economics while meeting data sovereignty and security requirements.

MongoDB does not provide equivalent functionality, and it is not available on AWS Outposts, Azure Stack, or GCP Anthos. Customers who wish to use MongoDB on-premises must license, install, manage, and secure MongoDB Enterprise Server on their own infrastructure in a do-it-yourself type of environment.

Does the service provide lower subscription pricing?

Autonomous JSON Database provides an average 3–5X lower subscription pricing than a comparable MongoDB service for the same memory, CPU, and storage, reducing a customer’s total cost of ownership. Autonomous Database Elastic Resource Pools can also provide significant cost savings.

MongoDB offers backup services for your MongoDB cluster, based on the size of the data files on disk. MongoDB charges $0.14/GB/month while Oracle Autonomous JSON charges $0.0244/GB/month, which is more than 5X cheaper than MongoDB.

Does the service offer true pay-per-use with per-second billing?

Autonomous JSON Database offers per-second billing with real-time up-and-down scaling of resource consumption to create a true pay-per-use pricing model.

MongoDB does not offer per-second billing, and autoscaling to smaller clusters occurs only once per day, leading to substantially higher costs.

Does the service provide cost savings via autonomous operations?

Autonomous JSON Database automates many tasks—provisioning, configuring, tuning, scaling, patching, encrypting, and repairing of databases—that significantly reduce the amount of effort to deploy and manage the database.

MongoDB requires much greater hands-on operations, resulting in higher administration costs.

Does the service provide cost savings through higher performance?