Item: Db2
Rating: 10
Author: Verified User

Overall Satisfaction with Db2

Use Cases and Deployment Scope

From the perspective of the engineer who designs the common infrastructure for our usage of Db2:

We use Db2 mostly for OLTP applications, some of them business critical with multi-million transactions per day
We also see increasing analytical workloads on these OLTP databases

Underpinning infrastructure is IaaS VMs built on RHEL and deployed on:

on-prem cloud platform (based on VMWare stacks)
public clouds like Azure, AWS, GCP

Pros and Cons

Pros

Rock stable database engine
Extremely consistent to deploy and operate across many major versions
Nevertheless increasing number of good features (e.g. time travelling, compression, encryption, Graph engine, etc.)
High level of compatibility with Db2 on zOS - easy to migrate

Cons

Hard to tell where Db2 really NEEDS improvement - I might have been working with that DBMS for too long...
Yes, there has been some reduction with Db2 support, but that's improved a lot again over the last year. And yes, there have been some flaws with new features and a few security vulnerabilities in the last 2 years, but that's also been stabilized quickly by Db2 development.
One thing that comes to my mind as an improvement would be a really great and flexible function for pivoting data. There are ways to achieve it, but flexibility and simplicity would be awesome.

Return on Investment

Compared to competitor products which are also strategic in our shop (i.e. the large global vendors), I see that the number of DBAs we need to manage Db2 is clearly smaller than for other RDBMSs.
As for ROI, Db2 has "always" been in use at our shop. All business cases with the focus of moving our existing Db2 LUW to a different RDBMS product (be it vendor or open source) resulted in clearly negative impact, i.e. high migration cost at zero additional value.
Particularly for applications which were migrated from Db2 zOS to Db2 LUW, the use of any different RDBMS product would result in massive additional cost to rewrite the mainframe applications which today continue to run against the Db2 LUW database mostly like they did before against Db2 zOS (embedded static SQL with packages, similar BIND options, security management, etc.)

Many different scalability options from the very simple and easy-to-manage single instance with 1 to many databases to pureScale clusters.
Plus of course the imho unique DPF setup which pushes Db2's MPP capabilities to stretch a single database across many (physical or virtual) servers and even extends such huge systems to support HA configurations.

Reliability and Availability

We've had exactly one P1 incident across all of our hundreds of Db2 databases in the last 8 years, and this one was caused by a SAN disk issue, i.e. Db2 was the victim.
We've seen some performance degradation on the verge of a major version upgrade which resulted in P2 incidents about 2 years ago. Needed IBM support and it took a few days until they provided us with a fix for the issue.
Other than that, Db2 just runs and does its job, 24x7, unexcitedly and without asking us to pat and caress it.

Key Insights

Do you think Db2 delivers good value for the price?

Yes

Are you happy with Db2's feature set?

Yes

Did Db2 live up to sales and marketing promises?

I wasn't involved with the selection/purchase process

Did implementation of Db2 go as expected?

Yes

Would you buy Db2 again?

Yes

Likelihood to Recommend

I haven't really come across a scenario for RDBMS where Db2 is not very well suited. But then, I design and implement large scale, highly standardized, enterprise-ready estates with the goal of cost-effectively operating thousands of databases. I.e. my scope is rather on the infrastructure side than on functional details of Db2.

Comments

Please log in to join the conversation

Db2 from the perspective of large scale infrastructure engineering and operations

Software Version