What is most valuable?
IBM dashDB’s availability on the cloud and integration with IBM Bluemix gives universal access to our teams that are spread over different geolocations. It’s a high-performance columnar database and stores data at 10x compression in comparison to traditional RDMBS systems.
For us, we have large volumes of data to store and analyze, i.e., around 2-10 GB daily data in flat file formats and with high compression ratio; thus we save a lot of space on the cloud.
IBM dashDB comes with some unique features like integrated RStudio, a full scale database management studio (IBM Data Studio) and web-based management console that makes using this product easy.
It also offers SMP as well as MPP instances where MPP is a 3-or-more node cluster for big data storage and analysis.
How has it helped my organization?
Our analytics teams really like the integrated RStudio feature with which they can instantly connect to structured data and do their analysis. Traditionally, our analytics team would ask for data in flat files. They had to write the codes in order to clean and import the files and only then could they start the analysis part.
Now, the ETL team takes care of importing the data to dashDB and our data science team can focus on building the models. Once the model is developed, our modeling team can save the results in dashDB, which can be instantly used by our presentation apps. This is much easier and faster to work with instead of using the CSV files in all steps of data sharing.
What needs improvement?
One of the biggest advantage of cloud computing is Auto-scaling. AWS Redshift allows you to add more storage and CPUs/instances without any significant downtime.
With dashDB, scalability and uptime need more improvement. dashDB is not auto-scalable as of now. For any addition of space or computing power, we have to raise a request and there is a downtime to upgrade the instance.
The pay-as-you-go model is missing. You must buy a minimum instance of CPU and storage to begin with.
For how long have I used the solution?
I have used this solution for more than two years.
What was my experience with deployment of the solution?
Deployment was smooth and easy. IBM's enterprise support is excellent in taking down the requirements, suggesting the right solution and planning the deployment as per your needs.
After our initial setup of single instance of dashDB with basic storage and computing, we moved to higher storage and also created multiple environments for Dev and Production. In each case, the planning, execution and support on dashDB services was very good.
What do I think about the stability of the solution?
In the first year, we had multiple issues with stability of the service. We started with a cloud-based shared instance with limited data space. Once our PoC was done, we moved to a dedicated instance with high availability and better SLAs.
What do I think about the scalability of the solution?
We did encounter issues with scalability. Even with a MPP instance, scalability is not transparent.
To add a node, we need to raise a request and there is downtime associated with it.
How are customer service and technical support?
For a dedicated instance, the technical support is extremely good. There’s a dedicated team to address any of the reported issues and usually the response time is under few minutes.
Which solution did I use previously and why did I switch?
We had data in MySQL and Oracle on AWS. We switched as we were looking for a long-term solution and also for data privacy issues.
Most of our clients have concerns for data privacy in the cloud. With dashDB dedicated instances in IBM SoftLayer infra, we can assure our clients of physical separation of their data.
Dedicated instances also offer dashDB in bare-metal servers. This gives us much higher performance in data computation and at a lower cost in comparison to virtual CPUs.
How was the initial setup?
It was very straightforward. You need to create a free account on Bluemix and start exploring dashDB along with the other services. Once we finalized dashDB, IBM took only two or three days to setup our dedicated instance.
What's my experience with pricing, setup cost, and licensing?
If you know your data is large to begin with and will grow quickly, dashDB pricing is in comparison to the other service providers in the market.
Which other solutions did I evaluate?
We evaluated a couple of other options. NoSQL databases and free databases such as MySQL and PostgreSQL. However, NoSQL did not work at all for us.
Our data was structured and relational; we needed something with the traditional cursor kind of behavior for our apps.
What other advice do I have?
If your need is for an integrated development environment such as RStudio, Spark, app development and other services as offered on Bluemix, then you can plan your data scalability well in advance.
dashDB is a good product to work with. The extra cost you spend on performance, technical support and tools to work with is worth it.