Maybe SAP customers got it all wrong. Maybe 2025 is not the deadline for moving away from AnyDB (DB2, Oracle and SQL server), but the year that Hana will finally be finished – final release 2025!
At the dawn of the Hana era, Hasso Plattner and former Chief Technology Officer Vishal Sikka promised customers Continuous Improvement. Version numbers for Hana were not even up for discussion. Famous German poet Bertolt Brecht would now say, “Yes, make a plan! It will just go up in smoke. And make yourself a second plan – neither one will work.”
And he would be right: several years ago, SAP presented the community with Hana 2. Now, every customizing project typically starts with the question: Hana 1 or 2?
People familiar with SAP’s database problems are already talking about Hana 3, expecting the new version to be released at the end of this year.
Why? Hana as in-memory computing database may be Hasso Plattner’s masterpiece, but it is also an eternal construction site. Numerous SAP service notes are proof of a very vibrant Hana development, with fatal error and kernel panic being almost standard by now.
A masterpiece was born – and destroyed
Everything started with the idea of a real-time ERP, a new ECC (ERP Central Component) which would be able to respond almost instantly. This real-time ERP/ECC primarily required a new database. Secondarily, SAP had to do away with many Abap tables – but that is a different story, namely the story of S/4.
Real-time ERP required an in-memory computing database, and Hana was born. Created at the Hasso Plattner Institute in Potsdam, Germany, and tested in IT labs in Walldorf, the concept proved to be genius. A new class of databases was born as SAP ventured into uncharted IT territory. A database revolution was on the horizon. And then SAP went and destroyed everything.
Hana was promised to SAP customers way too soon. Hana was still a prototype, not having outgrown its beta phase just yet, and SAP already finalized licensing contracts and tied its revenue goals to the database.
Even five years before SAP’s own deadline, Hana is far from marketable or enterprise ready. Of course, there are thousands of Hana installations, but they hardly ever work as they should, as the numerous SAP notes prove.
Can Persistent Memory save Hana?
Now, Intel and SAP created a new memory class for Hana: Persistent Memory. What’s revolutionary about it is that Persistent Memory is more cost-efficient and almost as fast as Dram, can store more information, and keeps said information stored even in the event of a blackout.
What does this mean? Future Hana servers keep information stored even without electricity, are more cost-efficient and have even bigger storage capacities. The best of the best!
However, let’s not get lost in exciting future possibilities. Hana servers with Persistent Memory are still in an early development stage. More comprehensive evaluations will soon follow, but it’s still way too early for mass production for the SAP community.
Better things will come. However, the community has no time to wait for these things, as 2025 is already looming on the horizon. Until that deadline, every customer should (has to?) have switched to Hana, if the database is ready or not.