VMworld 2012 San Francisco / Barcelona



Today was the first day of VMworld 2012 in Barcelona, Partner Day.

After a few lab sessions I attended breakout session PAR2172 – Built to last: Why VMware vCloud should be the corner stone of your (customer’s) Private Cloud by David James. The main question was : ‘What do you want to build your infrastructure on?‘.

If you want a reliable house/infrastructure, you want a solid foundation, the structure on top of the cornerstone/foundation determines the stability of the whole structure. These days many organizations undergo a business transformation through an IT transformation. The ultimate goal they are aiming for is a strong and reliable infrastructure that minizes risks, vulnerabilities and susceptibility.

Tale of two different approaches
The first approach is a reliable one, based on trust and a proven track record. The second approach is the one based on announcements 12 months in advance with a refresh-cycle of 2-3 months where the final feature set and delivery date is uncertain. It should be no surprise that the first one is the VMware approach and the second one is Microsoft. Recent example: on September 4th Microsoft released Windows Server 2012 Hyper-V but to manage this you will need Microsoft SCCM 2012 SP1 which is not available yet. Rumor has it that this will be released in Q1 2013, which leaves a 4-7 month timespan in which you cannot manage your Windows Server 2012 SP1.

In this session there where 4 areas of interest:

  • Virtualized Infrastructure Management;
  • High Availability & Disaster Recovery;
  • Resource Management;
  • Security & Networking.
Virtualized Infrastructure Management
With vCenter Server VMware offers a comprehensive management suite from a single tool. With the System Center suite Microsoft offers a management suite which consists of many different windows, snap-ins, pop-ups and scripts. Certainly no single pane of glass. This was exactly my response after the System Center session at Microsoft’s TechDays in Februari of this year. In the Virtualized Infrastructure Management space Microsoft lags a few years behind, I can agree with VMware on this.
High Availability & Disaster Recovery
With VMware the implementation of High Availability is purpose build for the VMware platform. For Disaster Recovery VMware uses Site Recovery Manager (SRM), again purpose build for the VMware platform. Disaster Recovery with SRM to a cloud provider eliminates the need for a dedicated DRS site. It offers application consistent replication and automated fail-back with vSphere replication which increases automation and reliability.
With Microsoft High Availability (MSCS) is a one size fits all solution which is based on physical infrastructure based Windows NT4 technology. So the same High Availability solution for file-, print-, SQL-, Exchange- and your Hyper-V environment. Microsoft does not have real Datacenter Disaster Recovery because SCCM 2012 lacks significant automation in critical areas. Their solution is based on scripting and manual tasks, so manual disaster recovery instead of fully automated.
In this area I think the VMware standpoint is a bit far fetched. Although Microsoft clustering is ancient, and in Windows NT4 very buggy, they improved quit a bit in Windows Server 2008. On the one size fits all argument, I’m no MSCS fan either but if it works, it works.
Resource Management
With resource management there’s an old model of IT management, manual with scripts and workflows, and a new model of IT management, fully automated and policy based. VMware is a supporter of the new model which gives you granular control of processor and memory resources through resource pools and easy pooling of resources (processor, memory, storage), configuration and view of allocated resources. Apparently with Microsoft SCCM 2012 you can only control processor and memory resources per single virtual machine. There’s no logical grouping of resources and no native support for viewing resource allocations. In this comparison VMware vCenter is far more powerful.
Than there’s the issue of Unified Resource Management. The issues with a multi-hypervisor solution are versatile:
  • Added cost and complexity;
  • ‘Breaks’ unified resource management;
  • Dual skill set needed;
  • Separate High Availability and Disaster Recovery needed.
So approach such a solution with ultimate caution.
Security & networking
Security is the number 1 concern within large companies but networking and security have not kept pace with the datacenter evolution. Networking and security are still handled as individual physical silos. With the whole vShield suite of products, VMware has a solution which fully integrates with VMware vSphere and vCloud Director. With Microsoft the network and security integration is immature, manual and relies on third party solutions.
So Microsoft has a hodgepodge of different management tools which are not purpose build to manage complex virtual infrastructures.
VMware has a reliable and complete solution which is proven, lightweight, adaptable, purpose build, fully integrated and automated and legacy free. Fundamentally the true cornerstone for a reliable private cloud solution. A statement I can totally relate to and agree with.