I just finished a panel discussion on the future of the data center with leaders from all four companies, and while they disagree on some practices, there was unanimous agreement that now is a good time to be consolidating servers, networking, storage, and even data centers.Riverbed wants you to consolidate your data centers. Servers in remote offices are expensive to maintain and a security risk for data. Cisco wants you to mend your silo'ed ways. It'll make systems that can be managed by networking, storage and server teams, but the real value is in breaking down those silos.

Brocade agrees, but reminds you that you've invested a ton of money in that Fibre Channel infrastructure, and there's no sense in totally junking what you've got. Consolidate, but utilize the investments and expertise that you currently have.

Juniper says virtualize it all, but watch out for proprietary technology that can lock you into a single vendor solution.

The future data center has a unified network for storage and server connectivity, it runs at least at 10 Gbps, and it's fully virtualized. The result is a highly responsive environment that takes far fewer resources to manage.

All the panel agreed on that. However, when asked about the management layer, and the orchestration of that highly dynamic environment, I was somewhat shocked to hear the usual suspects rattled off as the likely providers: Tivoli, OpenView, Unicenter, BMC and the like were the names brought up.

I suppose it could be that these will be the providers of the right orchestration tools, but for many organizations, these players didn't do an adequate job of managing the static environment of the pre-virtualized data center, what would make anyone think they'll be able to step up to the challenge offered by the next generation data center?

To some degree, this group of vendors is doing what they should be doing. They're reworking the plumbing in such a way to enable the consolidated virtualized data center of tomorrow, but they also need to step up to the management plate and become proactive proponents of good orchestration. Cisco's UCS lays down some of that groundwork, but at least for now, it's all done in a proprietary way.

More specifically, the Cicso / VMware Vswitch is good technology, but that it's limited to Cisco and VMware should probably be enough for most IT organizations to limit its use to testing, evaluation, and planning.

Juniper, Brocade, and Riverbed agree that an open Vswitch specification is what's needed, but can they pull it off in a meaningful way without the participation of Cisco? That's the drama that's ensuing. And until the orchestration picture is clearer, server and data center consolidation will continue in relatively small steps.