- Paul Burgess
Gartner's "10 Ways to Quickly Reduce IT Costs" review
On August 8, 2022, Gartner published a blog around reducing costs quickly. The summary is outlined in the graphic below.
If I think about the list and the broader blog, I would make this a baker's dozen of a list (that equals 13 for those who are not sure). As a student to thoughtful and defensible research I found a more measured source from a group called Platform 3 Solutions involving someone I have found as a great researcher and story teller - Tom Rieger. This collaboration of people have done a multi-year analysis (year 1, year 2 and year 3) and surveying of the market and have bubbled up additional cost cutting motions happening already that everyone should consider. What is different here is you can see change over time (or no change), the evolution of the research has grown and pivoted with the macro market and - most important - it is transparent and laid out for all to evaluate in details.
These are the 3 items I would add to this list:
#11 - Discover the next, best platform
The choices today for 'platforms' are confusing. Things like:
If cloud, which one?
If on-premise and how?
Containers/Kubernetes or VMs good enough?
In this context, 'best' leans into the least cost that also brings the levels of service you need for you business. Consider Amazon Aurora Postgres and how it only promises 99.99% uptime - that is almost 9 hours of downtime a year that would be mostly based on catastrophic failure events.
An executive for a Fortune 10 organization stated " Over 80% of my data processing is still in data centers and I can still run most workloads cheaper on-premise and maintain the uptimes my business needs". Charging into the cloud may not bring the greatest TCO savings and be careful as there are many opportunities to pay for components that bring zero value (like the high storage and IOPS settings).
#12 - Target technical and data debt
Decisions and systems written 20 years ago are now very expensive to maintain and cause organizations to be stuck in older business models. This is also not just about what is running (an old Oracle Exadata appliance with nearly end-of-life Oracle database), but also the data inside the application. As system run some of the largest tables in a database are not business related - but system, security and tracking related. Over time, these become some of the largest bloat of an application. The results are unhappy client experiences (both internal and external), slow and expensive systems and a platform stuck in the past.
#13 - Open Source stack is the way forward
Linux and Postgres have been opensource from their roots over 20 years ago. In that time, the open-source choices in the market exist for everything an IT shop needs. One of those last and most important pieces is the 'database'. Not here to debate NoSql vs SQL - but the best open-source database in the world for those more operational and mixed workloads is Postgres.
Born out of my where I taught - University of California at Berkley - Postgres is the purest of open-source with a large global community of people and organizations making it better. The DB-Engines website has awarded it 'top database' for 5 of the last 7 years, it is the primary DBaaS success story for Google, AWS and Microsoft Azure, and has the broadest set of ecosystem technologies and platforms in the world. In my research, organizations like EDB have also brought other complimentary tools and services to cover those top-tier workloads,
In a conversation with them, it is clear they have the focus, funding and fortitude to tackle any sized needs and organizations as they do now for the likes of American Express, Mastercard and other Fortune500 and government entities for their primary systems of record. They have a version of Postgres that makes the migration off Oracle academic, bring the most advanced Kubernetes offering and are quickly building out the most advanced DBaaS offering in the market.
The Baker's Dozen in summary
While Gartner is generally right around the top 10, consider adding the above mentioned as it brings actionable opportunities as most organizations are now in this mode of (1) what do they have left this year to spend/do and (2) plan for 2023.
Choose wisely the next platforms - Is on-prem still the best cost option and when choosing a cloud, do not get locked in.
Make the hard decision to migrate off that old data platform (generally running on Oracle, SQLServer and DB2) and onto Postgres. From the research, Oracle is the most likely to be retired in organizations.
Do not lift-and-shift everything as-is to the new platform. Clean out the old data as the migration happens. "Storage is cheap" but in the cloud it can quickly become a #1 cost.