History lesson ... In the documentary "Revolution OS," I distinctly remember Eric S Raymond commenting how he couldn't believe Linux got the database vendors. For me, among others who saw it happen, it wasn't lack of belief. There was a discrete reason reason why _all_ major vendors starting supporting Linux overnight (within 6 months of that discrete event).
The point-of-inflection for the PC was its peak CPU and I/O performance was the late '90s, not just price-performance. The last bastion of RISC/UNIX had been broken, and commodity PCs started meeting or exceeding even the mid-level RISC/UNIX platforms in peak CPU and I/O (short of going to a costly, high-end RISC/UNIX platform with very exotic system interconnects), while commanding up to a 10:1 price/performance. It was true for the database world and I/O as much as engineering and scientific fields. It started with the "bundling+rebate" approaches of Microsoft consumer products via PC OEMs in the '90s. MS Germany's Kempin proved it with Vobis. The strategy was then used to leverage the 90%+ control to increase Windows adoption, then Microsoft application marketshare, where the Microsoft product's marketshare was a minority. It worked because it was bundled with a Microsoft product that was the defacto standard option in over 90% of sales. So by 1998, Microsoft decided to adopt product bundling at the PC OEM into the Server relam as well. After all, the PC Server was taking over the SMBs and even enterprise departments, and considered viable against even many mid-level RISC/UNIX platforms. And if there was going to be a database on the PC Server, it might as well be MS SQL Server, right? So the agreements (of which only the PC OEMs can testify what those terms were) caused many PC OEMs to only ship MS SQL Server with Windows NT 4.0 Server. Within six (6) months of that happening, you saw Oracle, Sybase and other products on the PC Server ... on Linux ... from PC OEMs. Virtually every PC OEM -- from Dell to Gateway, not just HP and IBM -- was sporting Linux with these database products at every tradeshow. It caught Microsoft completely off-guard at the floors of these tradeshows. Microsoft literally help give Linux an unexpected jolt of enterprise adoption. And the rest is history. The case-in-point here is that if you cause enough pain for your customers, the customers will eventually find another solution via another product and/or distribution channel to ease their pain. And that option will very likely not include your solution at all, let alone may undermine your existing product marketshare. It may not happen overnight, and even MS Windows Server didn't initially suffer server sales losses or inhibited growth at the expense of Linux. Most early 21st century Linux adoption was often at the expense of RISC/UNIX sales. But that has no longer been the case for the last few years. ;) ----- Original Message ----- From: Brian Long (brilong) <bril...@cisco.com> Sent: Tuesday, February 21, 2012 3:12 PM Feels like Oracle is becoming what Microsoft used to be (or still is). "You must buy our OS if you want to run our database in a supported fashion." ... _______________________________________________ rhelv5-list mailing list rhelv5-list@redhat.com https://www.redhat.com/mailman/listinfo/rhelv5-list
