Vendor Management
One throat to choke
A DBA perspective..."One throat to choke" is one of those pithy aphorisms that sticks in the mind. It is, perhaps, a negative way to look at vendor relationships in general (hence alternatives like "one hand to shake" and "one back to pat" becoming more prominent).
When it comes to a support contract, "one throat to choke" can be a good way to go. If your Oracle database breaks, which do you prefer: interaction with a single support organisation who are accountable for the fix, or a back and forth between two (or more) support organisations all too keen to lay the accountability at someone elses door? When it comes to Oracle, it's why I prefer to run Oracle database on Oracle Linux. The moment you introduce another vendor into the stack you weaken your hand. You introduce an opportunity for unnecessary delay. You risk becoming the intermediator in the troubleshooting conversation.
But, also remember another pithy aphorism, "Don't put all your eggs in one basket". The further down the stack that the issue gets from your area of accountability, the more "one throat to choke" diminishes in value to you. Your Oracle database crashes; when Oracle Support suggest it could be OS configuration, it helps if there is also an OS support contract with them. Suppose they then confirm an OS issue. In most larger organisations, accountability for the OS sits with the sysadmins, not the DBAs. Ignoring the benefits of Oracle Engineered Systems for the purpose of this argument and assuming your database server is just a generic x86 pizza box... what is the benefit to the DBA of having an Oracle x86 pizza box instead of a Dell (or HP or Lenovo) x86 pizza box? It's possibly not zero, but does it outweigh the benefit to the sysadmins of having all their UNIX servers (including those that don't host Oracle databases) under a single support contract with someone else? (And might the people who negotiate the cost of that support contract be able to do a better job if they could move that support contract to someone else if the incumbant supplier put up their prices too much? Competition is good for buyers).
Sometimes it pays to put your eggs in one basket. Maybe the basket is simply better than the other baskets for your purpose. Maybe it's specifically tuned to carry eggs. Weigh up the benefits of Oracle Engineered Systems for databases vs hyperconverged systems for everything else... where do you compromise? Do you buy Oracle SuperClusters for everything else, just to have "one throat to choke"? (Oracle think you should). Do you run Oracle on VMWare (with all the complications that brings) in order to standardise on Dell vxRail? (Dell think you should). Do you mix and match and complicate the life of your sysadmins? (and maybe your DBAs.. how about OVM on vxRail?). Or, perhaps Cloud? Note also that it's not quite as complicated when you use Microsoft SQL Server instead of Oracle Database for your thought excercise.
So, what's my message... "one throat to choke" is an important consideration, but it's not the only consideration. Pithy aphorisms are no substitute for actual thought.