- Cassandra is very strong for saving the time series based transaction data model, simply by reversing the time series order when creating the data table, we can very quickly fetch the "latest" records even from millions of associated transactions because the latest record is always at the top of the search. By combining with the TTL feature of the Cassandra column, it is easy to "auto" delete the old data.
- Cassandra combines the key-value store from Amazon's DynamoDB with the column family data model from the Google's BigTable, which makes it easy to manage both structured and non-structured data model efficiently.
- By using the DataStax Enterprise version provided Solr integration, it can even solve some ad-hoc query needs which may not be fully taken into account at the beginning of the project when the data table is created. This extremely adds more room to play for a large enterprise or project which does require some flexibility in the practical context.
- The linear scalability provided by Cassandra, allowing us to easily scale up/down the cluster by simply adding/removing the servers.
- The throughput for both the read/write performance of Cassandra is quite good.
- Managing the big cluster of Cassandra , even with the DataStax Enterprise Version, is still quite challenging for a maintenance team, considering the frequent version upgrade (even in the rolling fashion) and more frequent auto-repair, for me on this area, a powerful tool should be provided to "automate" this process as much as possible.
- The TTL design is good, however the pain is if the TTL is set on some data already inserted, it can not be simply updated. Unless that data is reinserted again, this fact causes a lot of issues in case the business strategy is changed which requires the purge strategy to be updated also.
- As the nature of Cassandra is still Java based, the GC sometimes eats some performance, if Cassandra can allow using more non-Heap memory space, to reduce the GC efforts which will free more power on the hardware.
- The default indexing strategy for JSON formatted data in the DataStax's Solr integration is not available. At this moment we have to implement our own to support our JSON text stored. We extract the key field from our data which might be required to be ad-hoc searched, converting them into the JSON format (only one level Map), and save them into the Cassandra column. On top of that we want Solr to index the key of each token.
- Cassandra lot of API's ready available for map reducing queries (like materialized queries).
- Cassandra uses ring architecture approach, there is no master-slave approach (like HBase). If data is published on the node, the data will get synced with other nodes in the ring architecture, compared to HBase which has a dedicated master node to orchestrate the data into its slaves.
- Write Speed
- Multi Data Center Replication
- Tunable Consistency
- Integrates with JVM because it's written in Java
- Cassandra Query Language is a subset of SQL query (less learning curve)
- No Ad-Hoc Queries: Cassandra data storage layer is basically a key-value storage system. This means that you must "model" your data around the queries you want to surface, rather than around the structure of the data itself.
- There are no aggregations queries available in Cassandra.
- Not fit for transactional data.
- Multiple datacenter usage w/ little or no data loss
- Rebuild/repair of objects (tables) in the keyspaces, allow to ignore keyspaces to repair.
- Monitoring tool form opscenter support for Cassandra 3.x (or some other open source tool)
- UI browser type to view data (rather than csql)
- Cassandra can preform read/writes very quick
- Nodes in a ring will keep up to date by sharding information to each other
- Cassandra is well suited for scalable application needing keyspace storage
- Cassandra's query language is clunky, which is likely due to the nature of NoSQL.
- Lacking the ability to relate data between sets makes querying harder, but this again is the nature of NoSQL.
- As a Java based NoSQL database it has the greatest community and adoption. Coupled with great Apache Hadoop, Apache Spark and Solr integration and a strong tools ecosystem(unit tests, stress testing), it is a unbeatable combination!
- As a hybrid architecture based on masterless architecture as in DynamoDB and column family data model as in BigTable, it hits the bulls eye!
- It has best in class performance across different kinds of read/write/mixed workloads. It provides linear scalability which works for the best performance, lowest latency and highest throughput.
- Being a tunable consistency model enables you to have consistency as your platform/application needs.
- If configured correctly, there is no downtime and no data loss.These are key criterias on critical domains.
- Apache Cassandra is lacking in some features, which DataStax provides in the Enterprise version. For example, security and advanced tools like OpsCenter. These would be a great addition to open source Apache Cassandra.
- At times we noticed some versions had issues not known in advance, for example, LostNotificationError on repair of nodes. However steadily the newer releases have become better and more stable.
- Examples of DataStax native driver with Cassandra 2.1 can be improved, as it does not provide all scenarios one would need on production.
- If you prefer to work with an open source project and be hands on, Apache Cassandra is one of the best. However if you need a managed Cassandra like service where you do not even want to configure/deploy/backup/restack, a DynamoDB service would be more preferred.
- Cassandra is JVM based NoSQL, hence garbage collector tuning is a key aspect, Garbage collection in JDK 8 and G1GC garbage collector is better or configure ConcurrentMarkSweep(CMS) garbage collector in an optimum manner.
Apache Cassandra is a NoSQL database and well suited where you need highly available, linearly scalable, tunable consistency and high performance across varying workloads. It has worked well for our use cases, and I shared my experiences to use it effectively at the last Cassandra summit! http://bit.ly/1Ok56TK
It is a NoSQL database, finally you can tune it to be strongly consistent and successfully use it as such. However those are not usual patterns, as you negotiate on latency. It works well if you require that. If your use case needs strongly consistent environments with semantics of a relational database or if the use case needs a data warehouse, or if you need NoSQL with ACID transactions, Apache Cassandra may not be the optimum choice.
- Continuous availability: as a fully distributed database (no master nodes), we can update nodes with rolling restarts and accommodate minor outages without impacting our customer services.
- Linear scalability: for every unit of compute that you add, you get an equivalent unit of capacity. The same application can scale from a single developer's laptop to a web-scale service with billions of rows in a table.
- Amazing performance: if you design your data model correctly, bearing in mind the queries you need to answer, you can get answers in milliseconds.
- Time-series data: Cassandra excels at recording, processing, and retrieving time-series data. It's a simple matter to version everything and simply record what happens, rather than going back and editing things. Then, you can compute things from the recorded history.
- Cassandra is a poor choice for implementing application queues.
- NoSQL requires thinking differently, and can be challenging for people with strong relational database backgrounds to understand. The CQL language helps with this, but it pays to understand how the engine works under the hood. That said, the benefits outweigh the challenge of the learning curve!
- Database compactions and anti-entropy repair can be burdensome on a busy cluster. Significant improvements have been made in recent versions, but it remains as an operational challenge.
- Cassandra is highly scalable.
- It provides the flexibility to store data in any format. You can add column family dynamically as need by the application.
- One of the best noSQL solutions I've used so far.
- A better UI access for reading the data.
- More graphical information to understand how the data is being processed, system uptime/downtime, etc.
- I used Cassandra-cli for running quries but it is not very helpful when it returns a lot of results. If there was some way to improve the user queries, it would be great.
- Performant. In particular, write performance is very good. Recently, a lot of work to address the changing systems environment has been done to take advantage of areas like SSDs and very dense storage systems.
- Distributed system logic. Multiple data centers and other common network configurations like heterogeneous nodes are handled and exploited well.
- Community. Strong community with users and project contributors worldwide. The open-source and commercial software people work well together with sharing of lessons learned and improvements based on feedback.
- Operational tools. Would like to see continued work to improve the operational capability for large clusters and large amounts of data. For example, analyzing the on-disk files.
- Repair. Being able to run repair continuously and with greater control to avoid any spikes in resource use.
I have simulated a few real time running apps like Facebook and Uber where I have used RDMS and Cassandra, and checked the performance using Jmeter. It clearly shows that Cassandra boosts the performance over RDMS. One thing I find difficult in Cassandra is following the documents, which are not so understandable.
- Undoubtedly performance is an important reason
- We have not encountered a single point of failure
- Scalability of Cassandra is good which is the most important for the companies where demand is scaling day by day.
- Cassandra has a wide range of asynchronous jobs and background tasks that are not scheduled by the client, the execution can be eccentric.
- Because Cassandra is a key-value store, doing things like SUM, MIN, MAX, AVG and other aggregations are incredibly resource intensive if even possible to accomplish.
- I think querying options for retrieving data is very limited.
It serves as the storage layer in our home grown sensor analytics platforms that utilizes spark for the computation. We use it to store billions of samples of wearable sensor data that is collected in various studies and experiments.
- High Availability - we utilize the data replication features of Cassandra. This enables us to access our data even when several nodes have gone down
- Data Locality - our architecture combines Cassandra storage nodes and computation nodes in the same machine. This enables us to utilize data locality and limit expensive network IO to read data.
- Elasticity - Cassandra is a shared nothing architecture. Nodes can be added very easily and they discover the network topology. As soon as a node has joined the Cassandra ring, the data is redistributed among the existing nodes and streamed to it automatically.
- Cassandra runs on the JVM and therefor may require a lot of GC tuning for read/write intensive applications.
- Requires manual periodic maintenance - for example it is recommended to run a cleanup on a regular basis.
- There are a lot of knobs and buttons to configure the system. For many cases the default configuration will be sufficient, but if its not - you will need significant ramp up on the inner workings of Cassandra in order to effectively tune it.
It is well suited for storing immutable data as deletes are extremely inefficient. As such, it is well suited for data archive and deep storage.
It is less appropriate for OLAP as has limited aggregation and filtering abilities, and no grouping whatsoever.