The proof of thousands was to determine the limits of VM. Just where do you hit the end of the road. What are the constraints?
Hundreds is used when talking about real work being done. In the early years of z/Linux, about 5 years ago, the applications talked about being move to z/VM were rather light weight. Routers, Web servers, Samba, FTP servers etc. At that point the rule of thumb was 1 engine could support about 100 Linux servers (consolidating 100 Windows servers to 1 390 engine...no IFLs at that time). An engine was about 120 MIPS at that time. Then, the big stuff started to become available. DB2, Oracle, Websphere. One copy can take multiple IFL engines (366 MIPS each with a limit of 32 engines on a z/990). So, you don't want to talk to people about running thousands of Websphere systems, or thousands of databases. But, given the right mix of applications and hardware, hundreds of z/Linux machines can be done and be productive. All depends on what you need to run. Tom Duerbusch THD Consulting >>> [EMAIL PROTECTED] 07/20/05 11:09 AM >>> A question about Linux on z/VM and scalability. Early tests of Linux on 390 (Test Plan Charlie and Omega) demonstrated that thousands/tens of thousands of virtual Linux servers could physically be supported by z/VM (and I note with respect that Sine Nomine has a presence in this forum). These days IBM steps up to the plate in public for 'hundreds' of virtual servers (ibm.com), and Charlie/Omega aren't referred to as frequently. There is a big gap between hundreds and thousands. My question ... is zVM (with up to 16 engines) really capable of supporting thousands/ tens of thousands of servers in the real world, and IBM just being coy? Or is it really hundreds? If it is in the thousands, is this theoretical, or are there really customers using it at this level? For the purposes of this question, let us assume a 'server' is used by a small business of 10-20 employees with several hundreds of customers, supporting web site services and email (in a penguin colony with non-uniform resource demands). Vague, but hopefully a pointer to something useful. I recognise that this question is dependent on a wide range of variables. These would include the fact that the actual workloads encountered will matter a great deal. Possibly there are many opinions about this, and no one correct answer, but I am very interested in hearing perspectives from those with first hand experience. Apologies in advance for asking such an open ended question, or if this question has been tackled in depth previously. I could find no references in the FAQ or recent postings. If there is a more useful way of defining a standard small business workload that you are aware of, please comment using that as a base. Before you criticise my lack of precision on workload, I hope you can comment on the basic question (admittedly posed from a perspective that is less technically informed on z/VM than most participants in this forum). Personal disclosure: I am a former IBMer, from the business consulting group rather than hardware/zSeries. My interest is in the potential commercial application of Linux on zSeries, beyond server consolidation. If this isn't the correct forum for this question, please forgive in advance, and point me to a better forum. Glenn Nicholas Holipac. ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
