TrustRadius
Couchbase Data Platform Review 20 of 24
Couchbase Data Platform Review: "Sitting On A Couchbase All Day Is Great For Your (Mental) Health"
https://www.trustradius.com/nosql-databasesCouchbase Data PlatformUnspecified8.370101
Brant Burnett profile photo
April 06, 2018

Couchbase Data Platform Review: "Sitting On A Couchbase All Day Is Great For Your (Mental) Health"

Score 10 out of 101
Vetted Review
Verified User
Review Source

Overall Satisfaction with Couchbase Data Platform

CenterEdge Software has been using Couchbase for a variety of use cases for many years. We started as simple query cache in 2012, then moved to shopping cart persistence, and now it's our primary database for our cloud-based products. Couchbase provides an easy to use but powerful platform that is very scalable, highly available, and performant. This allows us to design an application that we can be confident will scale to handle all of our customers' needs as we continue to onboard accounts. The ability to avoid downtime to perform scaling is particularly important since our customers are distributed around the globe.
  • Scalability - As our needs grow we can scale horizontally, taking advantage of autosharding. This is much less cumbersome than dealing with manual partitioning in RDBMS systems.
  • Availability - Automatic replication and failover within the cluster means we have no downtime, even if a node fails. It also means that we can scale up or perform database upgrades with zero down time.
  • Agility - The schemaless nature of Couchbase allows us to continuously deploy new application versions that add additional data to records, without the overhead and complexity of using DDL to alter table schemas.
  • Query Optimization - While the SQL-based query language is very robust and powerful, the automatic query optimization isn't as mature as RDBMS platforms. This requires a bit more attention to indexing and index hints.
  • Management UI - The management UI is much better than most other NoSQL platforms, but the need to connect to different nodes in the cluster depending on which function you are performing (FTS queries, N1QL queries, view management) can be a minor frustration.
  • Index Replica Management - The process for managing index replicas for high availability is not yet a streamlined as the other services, and can require some manual maintenance.
  • N1QL's SQL-like syntax has made bringing new developers up to speed both easy and inexpensive.
  • The schema-less approach to data store has increased agility.
  • Operating at scale, Couchbase tends to be less expensive than major RDBMS solutions, especially once you implement high availability in RDBMS.
Clusters can be easily scaled from a single node for local machine development up to massive scale with dozens of servers. This can be done with complete transparency to the application, with no need to make software changes to support the larger cluster.
Couchbase's performance has been excellent in our experience. In one particular case, the use of Couchbase instead of RDBMS improved page load speeds by an order of magnitude. The fact that adding more nodes scales both the read and write performance automatically and transparently has been particularly advantageous for our OLTP use cases.
Flexible data modeling has really fit well with CenterEdge's agile development approach and our move towards continuous delivery. As product owners request new features that require persisting additional attributes, the process is completely frictionless. A developer simply adds the new attribute to the class in their code, changes the code to use it, and commits the code. This moves down the deployment pipeline all the way to production, with no need for creating and running DDL scripts at each phase. Even more complicated changes, such as change an attribute from a scalar value to an array of scalars, are vastly simplified.
At this point CenterEdge hasn't moved into the field of mobile application development, so we don't have any experience with Couchbase Mobile. However, we are investigating it as a possible method for providing an offline mode for on-premise desktop applications.
I found the Couchbase provided a more robust solution that was easier to use and develop against. DynamoDB is limited to Amazon, which would reduce our cloud provider flexibility in the future. MongoDB has great read performance, but write performance doesn't scale as well, which was a detriment to our OLTP use cases. Perhaps most importantly, the N1QL query language is by far the most powerful and the easiest to use amongst all NoSQL solutions I've investigated.
Couchbase is very well suited for any cloud-based deployment where performance, scalability, availability, and agility are valued over ACID transactions. This includes online stores, social media, IoT data streams, and much more. It also fits exceptionally well with microservices architectures, where eventual consistency (without ACID transactions) is the norm. Any development shop using Domain Driven Design techniques will also find the document storage approach of Couchbase fits well with aggregate persistence patterns as well.

Couchbase Data Platform Feature Ratings

Performance
10
Availability
9
Concurrency
8
Security
10
Scalability
10
Data model flexibility
10
Deployment model flexibility
10