VMworld Europe 2012 – General Session
(This article reflects the general session of Tuesday. It isn’t a literal transcript)
Before the first keynote starts the General session was opened with a spectacular show with dancers and music, etc.
First on stage is VMware’s Chief Marketing Officer Rick Jackson.
- It’s 11 degrees centigrade in Copenhagen. ‘We’ve made the right choice’ (It’s about 26 degrees centigrade in Barcelona right now)
- 8,000 attendees;
- VMUG – 75.000 members worldwide, Europe has 45 regional chapters.
Wednesday the general session will be about “Empowering the workforce of tomorrow, that’s here today” by Steve Harrod.
Rick introduces Pat Gelsinger
In 2008 25% of the workloads was virtualized, 60% in 2012 (depending on who you ask). More than 90% in the coming years? It would be nice if only real old systems, that belong in a museum, aren’t virtualized. The waves of change in IT are noticeable. From mainframe -> mini computer -> pc/micrporcessor -> to the current situation network/distributed computing / virtual/cloud computing. With it, IT shifts from reactive (packaged applications, flat IT tax, project-centric dedicated stacks), via pro-active (service catalog, cost & use metrics, dynamic resources pools), to innovate (new business applications, pay per use).
The time to provision servers was weeks in 2008, now it’s a matter of a few days or hours. In the future it should be possible in minutes or mere seconds. In 2008 server provisioning was hard work with a lot of physical things to do. Today deployment of servers still is based on the physical world: network settings, memory, firewalls etcetera. The future is a policy based software-defined datacenter. Software-defined datacenter is defined as “all infrastructure is virtualized and delivered as a service and the control of this datacenter is entirely operated and automated by software”
The path to improve the efficiency in the datacenter is: abstract -> pool -> automate. First virtualize, then group the resources in pools. After that automate the delivery of IT services. VMware’s management philosophy is: more automation, less management! Because of that, multi-platform, hybrid, multi-provider is a must. In traditional management each stack had its own management stack. Because of this business agility suffers big time. The future holds the next generation of management: Simple, automated management for the cloud. To be able to deliver this products are updated or added to the vCloud Suite.
- Service provisioning: vCloud Automation center 5.1 (new), vFabric Application Director 5.0 (update)
- Operation management: vCenter Operations Management Suite 5.6 (updated)
- Business management: IT Business Management Suite 7.5 (updated)
Organizations that are thinking about the cloud often have a lot of questions. Often they ask: ‘what technology should I use?’. The answer on the technology & architecture side is vCloud Suite and vCloud Architecture Toolkit. On the question of: ‘How do I operate in this new world?’ the answer is:
- on the process & Control side: operational readiness, cloud operations advisory
- on the people and organization side: organizational best practices, role-based certifications
- On the IT Management side: CIO score card, financial models
Pat mentioned the Cloud Ops Forum, consisting of CSC, Cap Gemini, Deloitte, Infosys, HP, VMware, Dell, Canopy, CGI, T-Systems and EMC.
VMware recognizes that we live in a multi-cloud world and that we need to address it as such. vCloud, support for physical environments, non-VMware hypervisors, public cloud providers are on the agenda.For PaaS they have CloudFoundry. For automated Service Provisioning they have DynamicOps. For software-defined networking & security they bought Nicira. VMware invested heavily into a heterogenous solution for the future with 2 billion dollar.
New applications and other uses like Big Data and HaDoop are changing the way we work with data. Approaching this from the same paradigma as always will give us the same results as we got until today. Rethinking solutions makes it possible to use the software-defined datacenter efficiently and manageable.
On the access side of things we have VMware View for the virtual desktops, but also Mirage for management of the physical desktops. With the combination, we can deliver a centralized Windows management solution that stretches across virtual and beyond to physical devices.
Pat introduces Steve Herrod – CTO
At VMworld 2012 in San Francisco VMware introduced the vCloud Suite, with a common install and a common look & feel. At its core you’ll find vSphere, that delivers the reliability, security and performance you need for your applications. 60% of workloads is already virtualized. The focus will be on virtualizing the business critical applications like Exchange, Sharepoint Server, SAP and Oracle applications. Besides the development this is about best practices and scaling guides.
Last year the mother of all VMs, the ‘Monster VM’ had 32 vCPU’s, 1TB memory and could do 1 million IOPS. 2012 delivers the “Monster VM on Steriods” with 64 vCPU, 1TB memory and more than 1 million IOPS per VM. All this will accelerate the virtualization of business critical applications.
For the near future the following things are in focus: telephony/VOIP, High Performance Computing, Big Data. This already lead to Project Serengeti. Serengeti is an open source project initiated by VMware to enable the rapid deployment of an Apache HaDoop cluster (HDFS, MapReduce, Pig, Hive, ..) on a virtual platform.
Until today the storage was aimed at LUNs. In the future the storage needs to now where a VM ‘lives’ to provide the right performance and storage at the right time at the right location. In order to do this you will see virtual volumes (vVols). Virtual Flash will also be something that is going to be used. With Virtual Flash all VMs can use storage on SSD, based on, for example, shares, limits, etc to offer tier 1 storage to all VMs. With virtual SAN technology local disks can be used to create a distributed storage solution. Adding an extra hosts will not only add CPU and memory to the cluster, but also storage to the solution.
Today application deployment is still a lot of work. Applications have a lot requirements: what network should I use, what type of storage can I use, VLANs, subnets, etc.etc. In the future a lot of things will be based on policies instead of hard manual labor. Deploying an application from vCloud Director also provides all the resources the application needs.
vCloud Director management is browser based. If you want to manage your infrastructure beneath it you can use ‘open in vSphere Web Client’ to deliver the single pane of glass management companies keep talking about. In 2012 native partner plugins will appear. One of the great things Steve mentioned was the multi-hypervisor approach. With plugins the functionality of other hypervisors can be managed from within the vSphere Client. The first version that is downloadable will be Hyper-V.
A short demo/screen for service provisioning
past: request, approve, provision: A lot of human activities in this.
now: request, approve, provision: All automated, driven by policies. This is done by the vCloud Automation Center (DynamicOps)
past: templates ‘ just a VM’: manual configuration/preperation, after that application installation
now: application blueprint, modular, automated; vFabric Application Director makes it possible to deploy services and applications based on blueprints. A lot of blueprints are already available in the Application Management Marketplace (beta). 3rd party vendors that already have blueprints available are Riverbed, Zend,Apache HaDoop
Before, operational management had limited visibility, manual analysis and limited knowledge. vCenter Operations offers an “Engineer in a Box” by providing best practices and explanations on errors or configuration issues. The last version had metrics at the VM level, now the metrics from within the VM can be used. VMware also introduces a dashboard for CIO’s, the ‘VMware IT Business Management Suite’. With the ITBM suite CIO’s can monitor and benchbark their costs. Couple of examples: cost breakdown, cost vs budget trends etc. Perhaps in the future it will even be possible to benchmark IT costs against other companies.
“Approaching problems in new ways”. VMware is always looking for a new edge in looking at your infrastructure. An engineer of VMware showed a very nice fusion between a virtual infrastructure and a social network (based on VMware Socialcast). You can approach virtual infrastructures the same way as a social network. If you consider resources and VMs as “friends” on Facebook you can monitor your infrastructure like you follow your friends. Each action has a reaction on the social network. The example shown was a failure (or shutdown) of an NFS server. This lead to a post on the news feed from the first machine. Other VMs experiencing the same outage reacted with a ‘like’ of this post. You could drill down to see which VMs are affected. I’m not sure if this will be in a product or not, but the idea is interesting. Thinking out of the box can lead to new approaches to century old problems.