On 3/26/07, John E. Conlon <[EMAIL PROTECTED]> wrote:
Enrique Rodriguez wrote: > ... > "With VMware's help, BEA ditches the operating system" (URL may wrap) > http://searchservervirtualization.techtarget.com/originalContent/0,289142,sid94_gci1234289,00.html > Sounds good, but will it be Open Source?
I doubt it. These particular vendors have a history of free versions, though.
> I've worked a decent amount with VMware and Xen running infrastructure > services. If not specifically any of the aforementioned vendors, then > at least this model (VM tight to the hardware) has a lot of merit; > partly for performance but mostly for management. > > I also think it has a big role to play in what we're trying to do at > the Directory project. For a long time I thought JNode was the answer > here, but this "operating system-less" model is even tighter. Why is so different between BEAs offering and JNode other than offering the ability of the running the BEA JVM alongside partitions with full OSes.
Besides differences in product maturity and performance, IMO the most important difference is the promise of their tools for managing large numbers of hosts/VMs.
In spite of the BEA offering instead of eliminating the JNODE I could see why the following three options could all still apply to ADS. 1. for a Dedicated appliance server = JNODE + ADS 2. for a Multipurpose Desktop = Native OS + JVM + ADS 3. for a Dedicated server = VMWare + BEA Liquid JVM + ADS From our stand point, being Java based do we care which of the 3 alternatives are chosen by our users?
From our standpoint as developers, we don't care. But, from my
perspective as a former Fortune 200 IT director, our ability to "play nice" with highly-desirable datacenter infrastructure is an important differentiator. This will be a selling-point that proprietary and other non-Java open-source solutions won't have. Enrique
