Here’s some notes from today’s Cisco, VMware and NetApps joint presentation on virtualization and cloud computing.
Killing the silos
It is critical that as IT managers that we understanding that silos in our organization can be both good or bad.
When looking at the typical IT environment, we can easily recognize that many of our applications have been siloed for historical reasons. You typically have barriers that prevent you from amalgamating your ERP with your HR and CRM systems. These application silos are not true business requirements, but more typically silos created by historical business processes and access requirements.
The first step we need to address when we look at managing our infrastructure costs downward is to break some of these silos and consolidate like applications. Do you really need a dedicated Oracle server for each application or can you amalgamate the instances onto a smaller number of physical devices and re-use those licenses to provide a higher DR/HA capability? Most organizations will see savings of 10-30% split relatively evenly between OPEX and CAPEX just by addressing these core business issues. More importantly is the released capacity and licensing that can mirror this reduction and address short term business growth without increasing your IT spend.
Once the total number of severs and silos has been consolidated you can effectively look at virtualization as a next step. Typical reductions you will see by introducing virtualization is again 20-50%, mostly in CAPEX but as much as 20% of that may be in OPEX due to reduced power/cooling and man power reductions. Note these saving again extend outside of IT and affect the business as a whole.
The challenge for larger organizations is the fact that silos can’t be easily torn down. This is where the concept of creating Zones of Virtualization comes in. One of the new capabilities that is possible with Vsphere, Cisco and NetApps is the ability to logically carve up a Virtual infrastructure environments to allow multi-tenancy. Essentially creating virtual silos within the virtual pool.
If that sounds familiar it is basically the historical concept of time sharing on a mainframe. It is also the basis of cloud computing and hosting.
The joint panel was branding this concept as IT-as-a-service (ITaaS) which is essentially an internal/hybrid cloud.
Since you can carve up and bundle your app/os/VM/networking and storage, you create a total computing environment package for the business area.
While most environments may not think they have an immediate need for this type of issolation and bundling, consider that these capabilities essentially create the ultimate DR/HA configuration. You can now fail, move or migrate a complete application environment between your datacenters or between your own dedicated datacenter and the cloud (private or public). That means by combining your DR budget/capabilities and your core computing you can often build a very strong business case for this level of capability on the cost of down time alone.
Most organizations can now easily achieve 95%+ virtualization in terms of their total server count. The story repeated by the panel, audience and all involved was we should make virtualization the default policy in or organizations and find the exceptions that can’t work well vs trying to prove it will work on a case by case basis.
There are loads, particular hardware configurations etc that will never virtualize. No one is countering that statement. The theme though is those are becoming few and far between.
Both clients and vendors are inherently moving away from these restrictive architechtures as the demand to do so increases. Realistically if you have one of these configurations, you need to carefully review the product against the costs of keeping them separate. If your organization has modernization as a core strategic component of your business strategy, you may need to challenge keeping these products that are holding you back.
New Support Offerings
The three hosting vendors are announcing the integration of their support offerings. You will now be able to call the vendor you feel is the core of the problem you are experiencing and they will own the call. If there is a spill over to one of the other vendors they will escalate and coordinate on your behalf. The idea is to ensure you are not punted between vendors and dealing with finger pointing. The three have created a cooperative support model where the first line support at each vendor will escalate to a Joint Escalation Team (JET).
The idea is if you have a communications issue between your VMware VM and your NetApps iSCSI storage and are using a Cisco switch architecture to bridge the systems, knowing who to call can be a concern. Under the new model you would call your VAR where your service contracts is hosted and they will contact the vendor the believe should own the call. At that point all 3 vendors will escalate through their three organizations simultaneously. This is more than a marketing point as the 3 companies have programs to cross train their internal support personnel on all three separate product lines so they can all effectively understand and communicate with each other. Each will have dedicated engineers to support each others products.
If we compare the uptake to virtualization and see it as a 8-10 year journey, true cloud (private or public) will be mainstream in less than 5. The technology has moved passed trying to address the SMB space and is now focusing on the national/multinational client set. The cool thing is this capability is bundled in your basic licensing and entry-level products that we all purchase as part of our renewals each and every day.
The joint support shows how important to these vendors virtualization has become.
Many thanks to Compugen for the invite to today’s event.