April 2015 Summaries

The Java SDK 2.2 release introduces a range of new features and enhancements aimed at improving performance and usability for developers. Notable updates include enhanced N1QL support with an extended DSL for index management and collection predicates, simplified retry handling through helper classes, and the introduction of an Entity Mapping feature that allows for easier mapping between domain objects and Couchbase documents. Additionally, the release includes API enhancements such as a method to check for the existence of a key without fetching document content and a reintroduced bulk document retrieval option in view queries. Dependency updates, such as the upgrade to RxJava 1.0.9, have also been incorporated to ensure compatibility and performance improvements. The development team encourages feedback on these new features, particularly the Entity Mapping functionality, as they continue to refine the SDK with more forthcoming features.

Using push and pull replication in continuous mode can be helpful when developing an application, but it can also have an impact on battery life. An alternative to continuous pull replication is using Google Cloud Messaging (GCM) to send push notifications to Android devices. To implement GCM syncing in an application, a new Google API project must be created and the device token retrieved from GCM. A wakeful service kicks off the GcmMessageHandler class, which performs a one-shot pull replication of data from Sync Gateway when a notification is received. The user profile document stores the device tokens, and changes to the database are monitored through the Sync Gateway Changes Feed endpoint, which returns a sorted list of changes made to documents in the database. When a change is detected, the GcmMessageHandler class sends a request to Google Cloud Messaging servers with the API Key and notification payload.

Couchbase is dedicated to fostering an inclusive and welcoming open-source community, emphasizing the importance of setting clear expectations through its Code of Conduct. The community aims to be open, empathetic, collaborative, pragmatic, and considerate, encouraging participation from individuals of diverse backgrounds and identities. Couchbase underscores the importance of resolving conflicts with empathy, collaborating transparently, and ensuring smooth transitions when members step down. The Diversity Statement highlights the commitment to treating everyone well, taking responsibility for mistakes, and respecting various dimensions of diversity. Any issues related to the Code of Conduct can be reported via email to [email protected].

N1QL features have been added to the Ruby SDK, allowing developers to work with structured queries in Couchbase. The library does not provide an ODM or high-level abstraction, but everything else is possible. To use N1QL, developers need to establish a connection to the bucket using `Couchbase.connect`, and then call the `query` method on the bucket object, passing a string in N1QL. The response from the server is returned as a hash with :rows and :meta keys. Developers can create primary indexes for their data using the `CREATE PRIMARY INDEX` query, and fetch documents by their keys or use conditions. To make examples simpler, developers can use functions to strip metadata fields and extract just values. The library raises a `Couchbase::Error::Query` exception if there is an issue with the query. Developers can explore the structure of the sample database further, answer practical questions, and even create custom queries using joins and conditions.

The new Couchbase Kafka connector enables direct streaming of events from Couchbase Server (version 3.0 or later) to Kafka, though it remains under development with an invitation for users to report any encountered issues. The project, inspired by a previous connector from PayPal, can be accessed on GitHub, with developer previews available from Couchbase's Maven repository and future general availability artifacts expected on Maven Central. Users can utilize a straightforward setup involving a filter class to pass only MutationMessage instances and an encoder class to convert document values into byte arrays, effectively establishing a Couchbase-Kafka bridge. The provided example demonstrates how to configure and run the connector using the specified Couchbase and Kafka server addresses, with instructions for setting up these servers via provisioning scripts.

The recent competition among leading NoSQL database vendors—DataStax, MongoDB, and Couchbase—has been highlighted by a series of benchmarks released over the past month. Each vendor's benchmark attempts to showcase their database's performance under specific configurations, but concerns have been raised about the transparency and accuracy of these tests. For instance, Couchbase's performance was potentially misrepresented due to cache misconfiguration, outdated client libraries, and unclear replication settings, while DataStax's Cassandra was criticized for disabling replication, impacting durability and write performance. Additionally, MongoDB's performance may not have been adequately represented due to the use of a release candidate and partitioning issues. The emphasis is on the need for transparent, real-world benchmarks that clearly communicate configurations to allow for reproducibility and fair comparison among the databases.

Recent competitive benchmarks sponsored by Couchbase and MongoDB highlight the importance of these evaluations in the growing NoSQL industry, especially as enterprises deploy NoSQL databases for mission-critical applications at scale. While it's crucial for benchmarks to be open, reproducible, and unbiased to provide useful insights for developers and operations engineers, they can legitimately focus on specific use cases that align with the sponsoring vendor's strengths. For instance, Couchbase emphasizes large-scale, mixed read/write workloads, whereas MongoDB targets smaller-scale applications. Fairness in benchmarking requires an apples-to-apples comparison and the use of recent software versions, while transparency involves clear communication of the measured use cases and open access for reproducibility. As the NoSQL market is expected to grow significantly, reaching $4.2 billion by 2020, these benchmarks serve as valuable tools for enterprises to select the most suitable products for their needs.