A long time ago, Oracle DB could handle workloads much, much larger than any of their competitors. If you needed Oracle, none of the others were even a possibility. There are even tales that it was a point of pride for some execs.
Then Oracle decided to put the screws to their customers. Since they had no competition, and their customers had deep pockets (otherwise they wouldn’t have had such large databases), they could gouge all they wanted. They even got new customers, because they had no competition.
Fast forward and there are now a number of meaningful competitors. But it’s not easy to switch to a different DB software, and there are a ton of experienced Oracle devs/DBAs out there. There are very few new projects built using Oracle, but the existing ones will live forever (think COBOL) and keep sucking down licensing fees.
VMware thinks they are similarly entrenched, and in some cases they’re right. But it’s not the simple hypervisor that everyone is talking about. That can easily be replaced by a dozen alternatives at the next refresh. Instead it’s the extended stack, the APIs and whatnot, that will require significant development work to switch to a new system.
To add a concrete example to this, I worked at a bank during a migration from a VMware operated private cloud (own data center) to OpenStack. In several years, the OpenStack cloud got designed, operationalised, tested and ready for production. In the following years some workloads moved to OpenStack. Most didn’t. 6 years after the beginning of the whole hullabaloo the bank cancelled the migration program and decided they’ll keep the VMware infrastructure intact and upgrade it. They began phasing out OpenStack. If you’re in North America, you know this bank. Broadcom can probably extract 1000% price increase and still run that DC in a decade.
It’s very much the Oracle model.
A long time ago, Oracle DB could handle workloads much, much larger than any of their competitors. If you needed Oracle, none of the others were even a possibility. There are even tales that it was a point of pride for some execs.
Then Oracle decided to put the screws to their customers. Since they had no competition, and their customers had deep pockets (otherwise they wouldn’t have had such large databases), they could gouge all they wanted. They even got new customers, because they had no competition.
Fast forward and there are now a number of meaningful competitors. But it’s not easy to switch to a different DB software, and there are a ton of experienced Oracle devs/DBAs out there. There are very few new projects built using Oracle, but the existing ones will live forever (think COBOL) and keep sucking down licensing fees.
VMware thinks they are similarly entrenched, and in some cases they’re right. But it’s not the simple hypervisor that everyone is talking about. That can easily be replaced by a dozen alternatives at the next refresh. Instead it’s the extended stack, the APIs and whatnot, that will require significant development work to switch to a new system.
To add a concrete example to this, I worked at a bank during a migration from a VMware operated private cloud (own data center) to OpenStack. In several years, the OpenStack cloud got designed, operationalised, tested and ready for production. In the following years some workloads moved to OpenStack. Most didn’t. 6 years after the beginning of the whole hullabaloo the bank cancelled the migration program and decided they’ll keep the VMware infrastructure intact and upgrade it. They began phasing out OpenStack. If you’re in North America, you know this bank. Broadcom can probably extract 1000% price increase and still run that DC in a decade.