Keynote 2 Video

Found it here:
mms://wm.microsoft.com/ms/msnse/0804/33036/MMS2008-Day2-Keynote-bradand.wmv

AMD's managing a heterogenous environment

Yep there is a theme here ;-)

AMD are all about the out of band management - Management that operates with hardware resources and components that are independent of OS control. Think like an iLo card in an HP Server.

They are part of the DMTF standards body that are working to make standards for these types of devices and the functionality they offer. This should ensure all the vendors products are interoperable. In this case the standard discussed is DASH - Desktop and Mobile Architecture for Systems Hardware. It leverages WS-Management which is another DMTF standard.

The kinds of things you can do are:

• Remote power on/off - wake for patching
• Remote boot
• Serial console redirect - change BIOS, watch boot sequence.
• HW and SW asset inventory - saves desktop visit if something fails.

Then some demos which were deathly dull, I have seen someone remotely power on/off a PC probably 5 times now!

Asset Inventory Service

I've seen a couple of sessions on this now so I shall summarise my thoughts.

Right now I'd say that this solution has only one place where we would consider it's use, and that's Royal Canin. Given the lack of network routing between the sites, this is a solution that would be able to deliver value very quickly (the installation would still need to be managed on a subsidiary by subsidiary basis though).

Microsoft have big plans for this product though, and some of the areas of interest are the software catalogue, which will be automatically referenced in Systems Center R2. To become a fully fledged service though, user authentication needs to be standards-based (I don't think the MS passport is an identity standard yet, but I'm sure Doug will correct me if I'm wrong) and the database will need to be accessible to enterprise business intelligence tools.

So apart from RC, I think this is one to look at again no earlier than version 2 (likely to be announced at Tech Ed later this year)

Analysing 12000+ IT Environments

This was a great session, full of lots onf information about how the Micrsoft Operations Framework and the Infrastructure optimisation surveys have helped companies do their deployments and move from Basic-Standardised-Rationalised-Dynamic.
There are some self assessments online
http://www.microsoft.com/technet/solutionaccelerators/cits/mo/mof/moftool.mspx
http://www.microsoft.com/business/peopleready/assessment/launch.mspx
however, when we met with the NYC Account team they said they would get us the full survey.

Further info on the Infrastructure optimisation can be found here > http://technet.microsoft.com/en-gb/infrastructure/default.aspx
They follow ITIL and COBIT too.

They have been running these surveys for 3 years and realised certain problems:

• Purchasing products does not always mean implementing
• Utilisation of products is usually around 0-10% of their feature set, some go up to 25% but virtually no one is ever above 50%. An obvious example the presenter said was Calendaring. All email has them and the ability to share, however the majority of employees are unaware/untrained etc.

The secret to service excellence is often pitched as people, process, technology. However it should read People+Process+Technology. Having one of two clearly does not make up for a lack elsewhere, the framework should help companies understand what they need to do.

There are 3 Infrastructure optimisation models that have been developed with Gartner and MIT. These are:

1. Core Infrastructure Optimisation Model,
2. Business Productivity Optimisation Model
3. Application Platform Optimisation Model

Each of the above have the four stages of development - Basic-Standardised-Rationalised-Dynamic.

A nice metaphor was used to describe businesses that are trying to progress their IT without dealing with each stage in turn - like a ship with 40 anchors, buying a bigger boat wont make you go faster.

As much as you try to go faster/ more dynamically by implementing new technologies, if you do not deal with the problems you are facing at each stage, all you are doing is introducing more complexity and instability and actually making this worse.

Their research has shown that if in any of the 3 optimisation models a business is still at the basic level, they will never be above standardised in the other 2.

It was interesting to ave a bit more of an in depth look at the MOF and IO models, I had only really skim read them before. It is certainly worth us having a chat about our areas and where we think we are and what that means. Maybe a team meeting topic?

Application Virtualisation & Streaming

I've attended a whole bunch of sessions covering different aspects of this topic, so I'll try to summarise my thoughts now we're done.

Firstly, the App Virtualisation product finally gives the Microsoft management suite proper user-targetting for software distribution, which it has always lacked. The main business benefit of the product is reducing the need for regression testing of different applications (i.e. conflict solver no longer required!). There is a secondary benefit in allowing previously incompatible applications co-exist, thus increasing flexibility/agility. This would be a great solution for COMPASS running Office 2007 alongside previous versions of Office.

The streaming aspect might be useful to deploy to users with limited connectivity, and there is clear potential to use this deployment mode to enable centralisation along with NGN V2. Lots more work to do here but also some big opportunities.

In summary App Virtualisation is something that should definitely be on our roadmap, but it's not going be a solution for everything.

System Centre Mobile Device Manager

I can't tell you just how bored I am of writing System Centre now
;-)

This lecture was about planning and deploying SCMDM. I thought it would be quite architectural but it was mostly run throughs of the wizards.

Key facts:

• Planning is very important as you need to touch many areas within IT - firewall, Active Directory and policies, security, SCCM and so on.
• Need to clearly define who 'owns' SCMDM due to the above.
• No Active Directory Schema changes are required
• Best Practices Analyser tool can be run pre and post deploymnet and it can report any errors or changes that need to be made.
• Device lifecycle is important to define as you don't wnat non-standard devices to cause problems.
• Coonsider who should support the end device and service, i.e. if the operator supports the phone but you support the service how are problems resolved where they interrelate? Need careful commercial negotiation before deployment.
• List of devices that will be Windows 6.1 can be found here > www.microsoft.com/windowsmobile/mobiledevicemanager/devices.mspx

Mat's thoughts on the 3rd keynote

I know Brian has already covered some of it, I just had some additional notes that I thought were interesting.

• Microsoft are beginning to use Data Centres in a container. They have their first one installed in Boulder, Colorado which provides Virtual Earth. It is entirely powered by wind power.
• They have found System Centre a great help with acquisitions because it can now monitor and manager diverse environments and is massively scalable.
• Average is 1 engineer to manager 5000 servers!
• Internal Microsoft IT are basically the beta testers for all the System Centre products.
• They deal with 1 million events per day!

I think it would be very interesting to spend time with their team to see how they do things. Same for the desktop area who are more advanced in terms of the newer products and their use. I'll blog that in another post.

Extending SCCM to heterogenous environments

This was a session presented by Quest to talk about their products that can further enhance SCCM. It is designed to improve the ability of SCCM into non windows areas and help achieve a better ROI.
These are branded Quest Management Extensions (QMX) and contain the following functionality:

• Support for non Windows OS - Red Hat, Suse (server and desktop), OSX, ESX, HPUX via agents.
• Agentless (reduced functionality) for Novell, Ubuntu and network devices such as printers, switches, routers and mainframes.
• Can add non windows OS into Active Directory.
• Patch and software distribution to the above (1st bullet).

QMX for SCCM 2007 are the above except network devices, if that functionality is required you also need to purchase QMX for Device Manager.

Virtual Desktops

A quick update on the Virtual Desktop session, which was presented by Pete Downing of Citrix. The solution is fairly similar to the VDI (Remote SDS) solution that is under development internally, with a couple of noteable exceptions.

The Citrix solution (demonstrated on top of Windows Sever 2008 running Hyper-V) uses a very shiny looking web interface for the end user, and the network transport between the client and the virtual machine is ICA rather than the RDP protocol that most other solutions use. Citrix claim a 15x performance improvement over RDP (maybe it should have read 1.5x) but given our experience with Remote SDS on RDP at Doane, where performance is not comparable to the existing Citrix terminal server solution (ICA transport again), maybe we should be looking at this solution more carefully.

Other nice parts of the solution include the automatic virtual machine provisioning which includes some scripting for personalisation, but until we can get to a stateless PC image I think we'll continue to struggle with deployment. At least it's given me something to think about!

Virtual Machine Manager

This session was all about the System Centre Virtual Machine Manager (SCVMM) product. Microsoft realise that the virtual world is driving a massive amount of growth for example, last year, 1 in 3 Enterprises bought Softgrid, they predict next year that will have increased to 1 in 2 of all Enterprises.

They consider that virtualisation should be a skill you learn as part of a techie skill set, not a specialism you struggle to achieve. SCVMM is here to help!

SCVMM is basically a pretty gui front end driving Powershell scripts, at the end of each wizard you can generally click a 'View Script' button should you want to see how things work or use it to make automated scheduled jobs for example.

P2V functionality is built in (costs extra with VMWare) as is V2V (convert VMWare machines to MS), although SSCVMM 08 will have the ability to natively manage both MS machines and VMWare machines. The first product on the market to take this holistic approach. The 2008 version will also have the ability patch offline machines.

There will be a delegation web portal too, this allows a user to login and manage their machine via a browser. This saves giving too many rights or having to deploy an application to that end user just to make life easier. (We deliver the VMWare machine application via Citrix). This is all controlled by Active Directory Group policies.

It is worth considering Server Management Suite Enterprise as this contains SCVMM, SCDPM, SCOM, SCCM - basically all the System Centre products.

W2K8 Enterprise edition allows you to run 4 virtual machines with no need to buy additional OS lics, Standard only 1, but Datacentre is unlimited.

Due to this Microsoft claim Hyper-V is 1/3 the cost of an equivalent VMWare environment.

One note of caution was raised with a nice quote from Gartner - Virtualising without good management is more dangerous than not using virtualisation in the first place!

Microsoft know how to party!

Seriously look at the rapture on the faces of the audience ;-)

The party also had Lan Halo, a chainsaw juggler, a stage for people playing rock band, free arcade games and so on. We stayed for a while but the rock, paper, scissors amazed us ;-)

NSV

I have noticed that mid afternoon is when the Mars snackfood comes out, rest assured I am eating as much of it as I can and even taking some away in case I get hungry later! If I were in the Great Escape I would be the 'Scrounger'!

No doubt there will be a spike in Las Vegas net sales volume this week due to my gluttony ;-)

I'm a bit behind with my posts due to spending an hour or so working our why I was going to appear on the bad boy SFAR exception list and making sure I did not. We have the Closing Party tonight and are just about to meet the NYC Metro Area account team from Microsoft so I'll try to catch the posts up tomorrow morning, so mid afternoon UK time.

Poll Added!

On the right hand side below the Tag Cloud. We'd like to know if you have been reading and if it has been useful. From my perspective I have probably retained a lot more information by taking notes and then writing them up and having the opportunity to think about what has been said.

I was hoping people would comment, perhaps with questions that we could have found out for you. It could be we are providing all the right information for you anyway?

If it has worked, then maybe as different people go to event they can do the same and help keep everyone involved.

Is anyone there?

It's our last day tomorrow, so last chance for you guys to have us field any of your questions with the experts here.

I'd also like to take the opportunity to solicit feedback on our blogging experiment but as Mat is uber-blogging admin I'll leave it to him to add the polling thingy.

How do they do that?

An interesting session for a couple of reasons. First, hearing about how exisiting customers are working around product limitations is a great way to find out the 'real' product limitations. Second, there was a discussion about 'best practice' application packaging and deployment which is as relevant to our current environment as it is to Microsoft System Center customers.

The slides are pretty good so I won't go into too much detail here, but some things we should think about are Desired Configuration Management (DCM) and a consistent hardware decommissioning process. Our process and organisation actually map fairly well to best practice with the noticeable exception of our insistence on 'bundling' the applications into monolithic updates, but it seems to me this stems from a lack of automation at the site end, so we should carefully consider what can be done to reduce the burden on our local infrastructure teams.

Where's the Party?

We were supposed to go to the Microsoft event on Tuesday night - we had tickets and everything, only to get to the venue and get told that the venue was 'at capacity'. Seems a bit odd to go to the trouble of ticketing an event and then distributing too many tickets, but what can you do?

What we did is to find a bar (can you guess which drink is Mat's?) and watch the world go by. We were assured that the bowl of suspicious looking green/white things were wasabi, but we chose to play it safe and order chunks of cow instead :-)

Last night we finally found a casino that had cheap enough tables for us to have a go at BlackJack (the thought of losing $50 a hand in the mainstream casinos was bit off-putting), and we spent an enjoyable couple of hours losing a few bucks, then winning it all back just to lose it some more. Overall we finished up, but there are plenty more days to go!

Third Keynote

So this morning's keynote was billed as "Reality of the Cloud", which was all about Microsoft's datacenters. You start to understand the scale of their datacenter operation really when they talk about 'piloting' a technology on 1,000 servers, and growth doubling from 10,000 to 20,000 server installations per month. The speaker was refreshingly honest and talked about recently starting to use the latest MS products to manage their environment as previous versions just didn't cut it.

It was also interesting to see the term 'Operation Excellence' in use, as well as her departmental key drivers being very similar to some of our principles (growth, efficiency and trust). I'm not sure MS are big on mutuality just yet though! Probably the most interesting thing she talked about was the impact on MS when acquiring companies, as they are often open-source, niche technology companies that are highly heterogenous. This has directly led to the interoperability features in the recent products.

In short an engaging keynote that drove home the scalability of the MS management products and an insight into the infrastructure that sits behind the scenes of services like hotmail and MSN.

Enterprise Desktop Virtualisation

This session was all about the "sixth" piece of MDOP. The Kidaro acquisition is not due to complete until May, so there are no release dates yet. They also confirmed that the pricing of MDOP will not change when this component is added in, always a surprise from Microsoft!

So onto the product itself, which is fairly clever. At a high level, it is basically a 'hidden' virtual PC installation running windows XP or Vista. Applications that are 'published' from the hidden virtual OS are presented seamlessly on the user's host PC. This could potentially be used to deliver an SDS XP experience on a non-standard workstation (e.g. contractor laptop). It's also the technology that was used to run the USB demo at the end of the keynote that Mat has already mentioned.

The product adds usability to the Virtual PC application in a number of ways. Firstly, the virtual machine is hidden, so the user experience is seamless. Updates to the VM image can be done on the fly, and the update mechanism promises to be very light on bandwidth as the agent checks if any of the updated components already exist on the host machine and only transfers the differences.

While it is very cool technology, I have to wonder about the horsepower required on the desktop to run this acceptably, and there are still a bunch of unanswered questions on how licensing will work.

We should definitely watch this space!