Orchestrator And VMM And Alice Springs, Oh My!

Orchestrator And VMM And Alice Springs, Oh My!

Our blogging competition winner Tommy Carron packed in a busy first day at TechEd North America 2013. The high points? The inner secrets of Orchestrator and systems designed to deal with lower bandwidth speeds. The low points? Diet Dr Pepper and Microsoft apparently deciding to place a data centre in Alice Springs.

After breakfast with the Lifehacker team, a New Orleans band entertained us on-stage. Then Microsoft’s Corp VP Brad Anderson arrived James Bond style in a sports coupe for the Keynote presentation. This lavish entrance segued into the topics of Windows 8.1 and the System Center R2 updates. The new Windows Server 2012 R2 feature called Workplace Join was demonstrated allowing users to join BYOD to corporate networks. This feature used automated phone callback for verification while segregating personal and corporate information on the device.

Microsoft Azure pricing improvements invoked enthusiasm, followed by the announcement of SQL Server 2014. The visually stunning Geoflow in Excel was demonstrated using attendee data, including real-time tweets, showing visualisations on a 3D world globe map.

Following the Keynote I chose to attend the Data Centre Transformation foundation session. Jeff Woolsey (Principal Program Manager) and Jeffrey Snover (Lead Architect Windows Server) drew an analogy with the early days of hardware, the configuration of DIP switches & setting interrupts, as compared with cloud computing today. Microsoft’s vision is to make cloud as easy as modern day plug and play hardware where configuration is transparent and just works. Interestingly Microsoft made the point that this applied to Linux as well as Microsoft’s Windows Server 2012.

One notable feature of the new R2 release is multi-tenant VPN. This feature lets you bridge between infrastructure, external service providers and Azure. Supposedly Azure’s automated storage handled over 50,000 network changes per day with managers only needing to define fabric and policy then letting automation do all the work.

It is also now possible to remote into a VM without network access to the VM using VMBus. That includes a VM that doesn’t have an OS installed. Other VM enhancements are live VHDX resize and live export cloning – an exact copy can be exported, say for test/dev purposes.

Hyper-V recovery manager lets you set the replication frequency which is useful for “customers in Australia who have satellite with ‘super dodgy’ connectivity” according to Jeff. I’m glad
they’re thinking of us down-under. This is especially important when it appears Microsoft thinks its forthcoming Australian data center will be in Alice Springs!

A demonstration using Virtual Machine Manager (VMM) provisioned a 2-node file server cluster from bare metal. VMM interrogated the target system’s hardware and generated a PowerShell script. which appears to be a very welcom norm across the System Center 2012 suite. A file share was deployed and its ACLs set – all from within VMM.

I thought the demo of the new Windows Server 2012 R2 enhancements of deduplication and tiering within Storage Pools was a little questionable. When dedup is on IOPS increases because of knowledge of where blocks are cached. It was a little bit of black magic I think I’ll need to consult the storage guys about further. Microsoft did made it clear that the feature isn’t about replacing existing enterprise storage, but to pool and provide tiering to unmanaged disks.

Michael Laeworth also covered the Windows Azure Pack that provides a consistent management API & GUI for the management of VMs, service bus and web sites. He demonstrated the creation of a ‘Hosting Plan’ that can be associated with web sites, VMs, SQL or MySQL services.

I then took a break for lunch in the massive food hall where I sampled a Diet Dr Pepper soft drink. Somehow I don’t think I’ll be trying any more of those 😉

Once my plate was clean I listened to Anders Bengtsson and Pete Zerger’s crash course in System Center 2012 – Orchestrator. They described Orchestrator as a tool that that provides an alternative to using standalone scripts and ad-hoc utilities that were created by somebody that nobody knows anything about. Quite a good summary! The GUI still resembles Opalis and hasn’t been updated to reflect new new System Center interface.

Orchestrator is a tool they said possible to understand in an afternoon and work with after a day. What can make things difficult is overly complicated runbooks. Anders explained that it is best practice to have only 5-10 activities in each runbook and to take a modular approach. This encourages reuse and makes them easier to read.

It is important to only use Orchestrator when it is the right tool for the job. When a runbook begins to execute software commands or perform inventory, you should be using Config Manager. They also stressed that automation with Orchestrator should only be considered when feasible both at the business and technical level. They explained using the ‘Flatten Returned Data Behaviour’ to simplify returning results and Loops for repetition. They finished with Service Center integration and said it brings value to Orchestrator even when it’s not your company’s service desk tool. The reason is that it can perform validation of your runbooks. Many of which can be found at the community site orchestrator.codeplex.com

I chose to attend the afternoon System Center Cloud Services Process Pack (CSPP) session. Focused on Service Manager, this promised to provide chargeback for enterprise scenarios. Interestingly the subject matter only attracted a scant 15 attendees. Despite this, there was more audience participation than the earlier jam-packed sessions, with several questions raised.

Ranganathan Srikanth explained that consumer behaviour is to over subscribe and under utilise virtual machine resources. He went on to say that traditional chargeback is cost recovery and that this burden is shifted to the service provider when in the cloud. Unfortunately self-service amplifies this. CSPP offers to address this by provisioning multiple tenants to multiple subscriptions with granular metering and chargeback.

Visit Lifehacker’s World of Servers Newsroom for all the latest news from TechEd North America 2013. And don’t forget: TechEd is coming to Australia in September. Click here for more information.