Second Source

For a long time, I resisted intensive centralism. My sense has always been for freedom, autonomy and individual responsibility. Unfortunately, economic advantages too often contradict a distributed organization.

A ridiculously small example, but indicative of our situation: In the canteen, I met a former tennis partner who now works for us in central IT purchasing. He told me about drive frames for our IBM and Fujitsu servers.

The frame accommodates 2.5-inch SAS/SATA drives and is made of a little metal, a lot of plastic, and a locking mechanism.

At our listed German distributor, one piece costs well over 100 Euros. A supplier from San Francisco offers exactly the same part for not quite 20 dollars.

This source was found by an apprentice in purchasing via Amazon.

These are nice anecdotes, but it triggered a thought process for me: We have a redundant IT infrastructure with multiple power, server, storage and diesel suppliers; last for the emergency generators.

The list of our second and third party suppliers would easily fill an entire E-3 magazine.

Failure safety and disaster plans are a matter of course in our global IT architecture. And by its very nature, it is not an issue that concerns IT alone.

Plant security, security, plant managers and many others are combined here in disaster recovery teams. No plan exists for the failure of a singular cloud platform like SuccessFactors, Ariba, Concur, HEC or HCP!

The situation is more complex than it might seem at first glance: You can operate a redundant data center architecture worldwide that even responds in "real time". But if there is a bug in the software that causes the algorithm to deadlock, then no failsafe hardware will help.

Naturally, there are also solutions for this, but they cannot be organized and financed in the normal commercial environment: A task is worked out by two independent programming teams.

The probability that both teams develop exactly the same algorithm and make exactly the same mistakes in it is very low. But even this double effort does not solve all problems.

In case one algorithm collapses, the other one hopefully continues to work - and you are sufficiently protected. But what if both programs don't crash, but give different results? Then you would need a third program to approximate what the correct final result might be.

Unfortunately, I cannot present a solution for the SAP community here and can only reflect some thoughts from our internal discussions: Given the current technical development and feasibility, the complete dependency of an existing customer through software-as-a-service, as represented by HEC, HCP and many SAP subsidiaries, does not seem justifiable. An IoT application with HCP seems feasible.

But what happens in the K-case? Where are the backups of the data? After a download: Can the data be processed further? In which format is it then available? What software is available locally and on-premise to read the data?

Naturally, there are good reasons for cloud solutions such as Business ByDesign and HEC or S/4 in the cloud. Those who need no IT infrastructure, few resources and a solution immediately are probably well advised with these instant solutions: Cloud computing also has a right to life!

But if you have been a loyal SAP customer for 30 years, have accumulated a lot of knowledge and experience, and can call the necessary ERP architecture your own, you will give SaaS models a wide berth. Hybrid clouds and hyper-converged infrastructure yes, but cloud computing with SuccessFactors, Hybris, Concur, etc. probably not.

My many years of professional experience show me that many problems can be solved from a different perspective. We will take our time and rethink the second-source problem for software. But in the same way, our SAP should also give or allow itself a little more time.

There are still scare stories about Hana and in-memory computing anomalies. The computer scientist is not surprised, because good software goes through a maturing process.

To this day, one of the biggest mysteries in the SAP community is why Hana is being pushed into the market with a crowbar. Pushing the deadline from 2025 to 2030 and allowing S/4 Finance and Logistics the necessary reviews is not a sign of weakness, but of sovereignty and foresight.

As I write these lines, the TV news brings the news that VW also does not have a second source for selected Golf parts.

I could not imagine a car manufacturer without a second source supplier until now - obviously there is. SAP already had to learn the hard way that there is no second source for SuccessFactors.