Another day of playing with SharePoint which varied between quite interesting and extremely tedious, I thought I’d share a little tip that my colleague, Alan Dodd, gave me tonight.
After copying a load of files to SharePoint in Explorer view, it’s a lot quicker to check them in using SharePoint Designer which, incidentally, is now a free download (although registration is required).
In my recent blog post about the R2 wave of Microsoft Virtualisation products, I mentioned the management products too – including System Center Operations Manager (SCOM) 2007 R2 (which, at the time was at release candidate stage). That changes today, as Microsoft will announce that SCOM 2007 R2 has been released to manufacturing.
New features in SCOM 2007 R2 include:
“Enhanced application performance and availability across platforms in the datacenter through cross platform monitoring, delivering an integrated experience for discovery and management of systems and their workloads, whether Windows, UNIX or Linux. Enhanced performance management of applications in the datacenter with service level monitoring, delivering the ability to granularly define service level objectives that can be targeted against the different components that comprise an IT service. Increased speed of access to monitoring information and functionality with user interface improvements and simplified management pack authoring. Examples include an enhanced console performance and dramatically improved monitoring scalability (e.g., over 1000 URLs can be validated per agent, allowing scaling to the largest of web-based workloads).”
Further details are available on the Microsoft website.
Jeff Atwood has a great blog called Coding Horror (I’ve written about it here before) and last year he (and some friends) started a new site for developers called Stack Overflow.
Stack Overflow is a Q&A site – but (and I thought this before I read it on Jeff’s blog post announcing Stack Overflow), it’s not like Experts Exchange – the site that charges money for access to community-generated content (thank goodness for Google’s cache) because it’s free – as community-generated content should be. Of course, not everyone will agree with my opinion there – I’m sure the people at Experts Exchange think they have a fine business model but I think it stinks to charge money for something that has been generated by your users. Anyway, back to the point, Stack Overflow is a sort of forum-meets-wiki-meets-blog-meets-digg site for software engineers to ask and answer questions, earning reputation points based on the value of their input.
So Stack Overflow is a great site for developers but I’m an IT admin-type… is there something similar for us? Well, no – not really. There hasn’t been, but 
now the Stack Overflow guys are launching Server Fault. You can hear more about it in Jeff Atwood’s recent interview on RunAs Radio but it’s aimed at IT professionals and system administrators, running on the same concepts as Stack Overflow.
I wish Jeff and his cohorts all the best with Server Fault and I plan to spend some time over there myself. Right now the site is in semi-public beta – if you have an OpenID and you know the password then you can get in – and the full public beta is expected to commence next week.
Together with my colleague, Alan Dodd, I’ve spent a considerable amount of time this week cleansing data in a SharePoint list. Quite why I can’t import e-mail addresses from an Excel spreadsheet to a column of type “Person or Group” I don’t know, but we couldn’t seem to make it work…
Checking names or browsing the directory was tediously slow; then I found that if I pasted the e-mail address, rather than the person’s name, SharePoint immediately resolved that address to a name and let me move on to the next record… might be a timesaver for someone?
A few weeks ago, I caught one of the IT Manager series of webcasts that Microsoft is running, where Andrew Fryer was introducing Business Intelligence for IT Managers (I’ll steer clear of the obvious joke about IT managers and intelligence there… I might want a job as an IT Manager one day…). This was an interesting presentation for a couple of reasons: it’s not a topic that I know well; and Andrew presented 290 slides in less than an hour (which sounds a lot, but it wasn’t – he used slides with just a few words or a picture, in rapid succession – and I like that style).
I can’t find the recorded version of the presentation online but this blog post attempts to capture the main points.
According to Wikipedia, Business Intelligence (BI) can be defined as follows:
“Business intelligence (BI) refers to skills, technologies, applications and practices used to help a business acquire a better understanding of its commercial context. Business intelligence may also refer to the collected information itself.”
That’s a bit of a mouthful, but basically it makes BI seem hard. So, let’s think about intelligence without the business context – is it: knowledge and understanding (we used to think the world was flat!); about meaning and context (some information can seem obvious without context); about foresight (to predict future events); the ability to solve complex problems; or the ability to make decisions?
We make decisions all the time – and some of those decisions are poor ones – so if intelligence is about making decisions (and decisions require people), what makes a good decision? The answer is information. The information provides the meaning and context together with the background information (knowledge and understanding, likelihood of a given event occurring, etc.) to allow someone to make the right decision.
Information has been used throughout history to share stories, to learn and to discover things. Over time, information has helped to provide answers and to unlock secrets, which allowed innovation. Information has provided answers – and answers have allowed people to make decisions.
In a business context, the information is derived from data (about people, products, places, etc.). Where there are questions (which products are best? how are the sales figures looking? how are my people?), some insight is required to provide meaning and to convert raw data to information.
That data needs to be stored – originally it was stored in paper files and later on computer disks and tapes – but it also needs to be managed. The advent of databases provided a means for data to be stored and managed but business applications were needed to abstract the management of the database from business end users. These business applications provided a better way to collect data, easing the process of data entry and managing access, to ensure that those who needed access were able to find answers to business questions. But is wasn’t easy – the data was sourced from many locations. Reports were one approach, but they were one-dimensional and siloed, fragmented and lacking insight.
The advent of data warehouses allowed data from multiple locations to be organised, managed and accessed – or consumed. Now the business applications could analyse as well as report and the term “Business Intelligence” was born. Promising more access, from more locations, BI vendors created demand for more data. This led to businesses wanting faster access to data (improved performance). But as the volume of data exploded, so did the use of personal computers, and most data ended up in desktop productivity applications like Microsoft Excel, and Access. There was no single version of the truth upon which to base decisions, the data was hard to maintain and the BI tools cost a lot of money, so vendors had to find a way to reduce costs and offer increased functionality. The result was a period of vendor consolidation in the BI tools space and the formation of a few BI platforms, from companies like Oracle, SAP, IBM and Microsoft, offering more tools and more functionality, for both online and offline access.
But, for all the promises, BI tools were still not working. Business users were still confused and the business couldn’t get the answers that were needed. It wasn’t about people – it was still about disparate systems, with access to data controlled by the IT department. An overstretched IT department. So business users started to circumvent IT departments, but the BI tools were not intuitive and the users didn’t have the time to be IT administrators. Suddenly BI was about usability, and turning data into the right format to be easily consumed by people, with that data managed by IT.
There’s not just the data in the databases to consider either – there is unstructured data too. That unstructured data comes from places like blogs, wikis, e-mail messages, documents, presentations, and videos – at one point analysts considered that 80% of business was conducted based on unstructured data.
So BI is about the right person, accessing the right data, at the right time – and it needs to be people-centric because it’s generally people that make decisions, not computers. Businesses need to do more to collaborate, search, and communicate about questions and answers in order to drive innovation. Even in today’s times of economic uncertainly, BI is still a priority at CxO level in order for businesses to do more with less, to provide better insight for better decision-making, for more people.
Reporting and scorecards are important components of the BI toolset, along with dashboards to display key performance indicators, for analysis. On the desktop we still use applications like Excel but the data lives in the warehouse. Other BI features include data mining (e.g. the shopping basket analysis that supermarket chains carry out using our loyalty cards). For unstructured data, we have portals for collaboration.
In today’s BI implementations the critical success factors are sponsorship (at a senior level in the company), a compelling need, a culture of analysis (rather than looking for divine inspiration) and, most importantly, partnership between the IT department and business users.
I don’t pretend to know anything about any of the specialist BI tools but, on the Microsoft infrastructure side, we already have some useful tools. Office gives us desktop applications like Excel, there are collaboration services in the form of SharePoint products and technologies, and we have a scalable database engine in SQL Server – there’s more information on Microsoft’s BI blog and learn more about the products on Microsoft’s BI website. There’s also advice on planning for BI in the SharePoint Server TechCenter, webcasts, videos, virtual labs and podcasts and more advice for IT Managers and their teams on the TechNet website. Finally, if you just want the highlights and a bit of technical analysis, Andrew Fryer’s “Insufficient Data” blog is worth a read.
After a successful joint meeting in March, the Windows Server UK User Group (including the associated LinkedIn group) and the Active Directory UK User Group are meeting up again, and this time the Vista Squad are joining the party too as we spend an the evening of 28 May 2009 at Microsoft’s London offices looking at Windows 7 (client and server).
This is the first Windows 7 event of it’s kind in the UK and will include talks on: Application Compatability for Windows 7 and Windows Server 2008 R2; Top 10 Reasons to Upgrade to Windows Server 2008 R2; and Enterprise Features in Windows 7 and Windows Server 2008 R2.
Check out the event website for full session and registration details.
For the last couple of weeks the news sites have been full of speculation and gossip about what is now referred to as Windows 7 XP Mode and Windows Virtual PC. Most of the reporting so far has focused on the high level concepts and, as the beta went live to millions of Windows 7 release candidate testers, this post attempts to give a little more detail about how Windows XP Mode works.
Before diving in to the technology, let’s have a look at why Microsoft felt the need to provide this functionality. Their vision is to drive the overall adoption of Windows 7 by eliminating legacy application compatibility issues in the enterprise, mid-market, and small-medium business sectors.
This is not a developer workstation solution; nor is it for consumers. It’s basically providing functionality to run legacy applications “seamlessly”, meaning that a typical end user will be unaware that their application is actually running virtualised. It draws heavily from MED-V in that IT administrators create a pre-configured virtual machine with a legacy operating system and applications to run isolated from the host operating system; however, unlike Microsoft Enterprise Desktop Virtualization (MED-V), which remains part of the Microsoft Desktop Optimisation Pack, Windows XP Mode will be more broadly available. In an interview with Mary Jo Foley, Microsoft’s Scott Woodgate gave the following description as to the differences between the two products:
Top-level answer:
- Windows XP Mode is designed to support SMB customers who do not use management infrastructure and need to run Windows XP applications on their windows 7 desktops.
- MED-V is designed for larger organizations who use management infrastructure and need to deploy a virtual Windows XP environment on Windows Vista or Windows 7 desktops.
He then continued with the following details:
| Windows 7 XP Mode with Windows Virtual PC | Microsoft Enterprise Desktop Virtualization (MED-V) |
|---|---|
| Designed to help small businesses users to run their Windows XP applications on their Windows 7 desktop. | Designed for IT Professionals. |
| Available as part of Windows 7 Professional, Enterprise and Ultimate Editions. | Enables Virtual PC deployment in larger organizations. |
| Enables users to launch many older applications seamlessly in a virtual Windows XP environment from the Windows 7 start menu. | Provides important centralized management, policy-based provisioning and virtual image delivery to reduce the cost of Virtual PC deployment. |
| Includes support for USB devices and is based on a new core that includes multi-threading support. | Is part of the Microsoft Desktop Optimization Pack (MDOP) |
| Is best experienced on new PCs from OEMs but will also be available for customers as a separate download. | v1 builds on Microsoft Virtual PC 2007 to help enterprises with their upgrade to Windows Vista when applications are not yet compatible. v2 will add support for enterprises upgrading to Windows 7 (both 32 bit and 64bit) and will support Windows Virtual PC on Windows 7. |
| v2 beta will be available within 90 days of Windows 7 GA. |
More information may be found in Microsoft’s Windows XP Mode press release.
To enable Windows XP Mode, Microsoft has produced a new version of Virtual PC – Windows Virtual PC (VPC) 7 – a client-side virtualisation product that runs on Windows 7 (32- or 64-bit versions). As Jason Perlow describes, it’s not using a type-1 (native/bare metal) hypervisor like Hyper-V (sadly, as if there were to be a client side virtualisation product based on Hyper-V it would be great for developer workstations) but instead uses a type-2 hypervisor (hosted) model. Unlike previous versions of Virtual PC though, the new version requires hardware assisted virtualisation capabilities (AMD-V or Intel VT), which are prevalent in many recently-purchased PCs (even if switched off in the BIOS).
Officially, VPC7 only supports Windows XP, Vista and 7 guests but, just like earlier versions of Virtual PC, there is the option of using emulated hardware so it’s still possible to run other operating systems – it’s just not supported. It’s also worth noting that not all Windows Vista and 7 SKUs are supported in a virtual environment. Something else that’s not supported is the ability to run 64-bit or multi-processor guest operating systems, nor is snapshotting. And, because the virtualisation components are incompatible, there’s still no support for moving virtual machines between Hyper-V and VPC7 either. I’ve already been fairly vocal in my feedback to the product team on this; their response is that the priority is on application compatibility (and, on that basis, I can see the reasons for concentrating on single-processor 32-bit Windows XP support) but continuing to maintain incompatibility between client and server virtualisation platforms seems a little strange to me.
VPC7 features include:
From a technical standpoint, there are three main VPC7 processes to be aware of:
vpc.exe is the base process for Virtual PC functionality.vmsal.exe is the seamless application launcher, which waits for a new application request and launches it. Once the application is closed it sets a timer before saving the VM state and exiting. This means that, when the application is closed, the virtual machine is kept up for a few minutes in case the user launches an application that requires it but after a short while it will be put into a saved state. In addition, because undo disk settings are managed within the virtual machine settings, logging off/shutting down/hibernation is handled as normal, with no virtual machine prompts about undo disks and saving state.vmwindow.exe is launched when VPC is not running in integrated mode.VPC7 will not run on the Windows 7 beta (build 7000) as it requires the RC (or one of the interim builds – I’ve seen it running on builds 7057 and 7068). I haven’t tried this personally but I’m told that it cannot be installed on Windows Server 2008 R2 either; however something similar is possible with Hyper-V by installing the Terminal Services Remote Applications Integrated Locally (RAIL) components (RemoteApp). Certain Windows 7 editions will include the Windows XP virtual machine, so there is no requirement to build a separate Windows XP image.
Architecturally, VPC looks like a hybrid between Virtual Server and Hyper-V: it uses the Virtual Server engine, including a scriptable COM interface for VM automation (and the security model has been modified so it can be called from PowerShell without needing to make security interoperability tweaks); it also uses the VSP/VSC/VPCBus model from Hyper-V; and it integrates RAIL components from Terminal Services but, because the Terminal Services technologies for integrated applications and enhanced desktop support run over the VPCBus, connectivity is available even if there is no network communication between the guest and the host. Because it’s built on the Virtual Server/Virtual PC codebase, VPC7 is limited to 4GB of RAM and 128GB VHDs.
Windows 7 XP Mode and Windows Virtual PC form a neat solution for application compatibility in Windows 7, drawing on established MED-V (from the Kidaro acquisition), Terminal Services (through partnership with Citrix) and Virtual PC/Virtual Server (formerly from Connectix) technologies. They are very much a point solution for application compatibility though and Microsoft still does not have a decent client-side virtualisation solution for high-end users (developer workstations, IT professionals with several desktop variants, etc.). Whether this is enough to allay concerns from Microsoft’s customers who baulked at a move to Vista as a result of the application compatibility issues is yet to be seen but with the general perception of Windows 7 riding high, this might be just the insurance policy that IT managers want to ensure that legacy applications continue to function. My main concerns with this solution are support (Windows XP is still end of life – and legacy applications may not be supported in a virtual environment), the overall complexity of the solution (however much it’s hidden from the end user, there are still two operating systems running on the hardware) and performance (virtualisation typically requires increased memory and CPU requirements – together with the need for hardware assisted virtualisation, this is certainly not a solution for legacy PCs). Whatever the situation, I’m sure there will be plenty more written on this topic over the coming months.
A couple of weeks back, I wrote about some of the new features in Windows Server 2008 R2 but I did say that was only part 1 as there were a few surprises in store (held back for discussion at TechEd this week):
If Windows Server 2008 was good, R2 is looking better. The release candidate is available now, with general availability expected in the second half of 2009 (although not confirmed by Microsoft on any official sites).
Earlier this week, I wrote an introduction to System Center Data Protection Manager (SCDPM). In that post, I mentioned that SCDPM 2007 SP1 supports Virtual Server 2005 R2 and Hyper-V virtualisation platforms but I’d like to elaborate on that and highlight the need to consider where best to deploy the SCDPM agents.
With SCDPM 2007 SP1, we can back up Windows and non-Windows guest operating systems on either Virtual Server 2005 R2, Windows Server 2008 with the Hyper-V role enabled, or Hyper-V Server 2008. Depending upon the guest operating system, it will either be a VSS capable or a non-VSS capable guest .
Linux, Windows NT, Windows 2000, Oracle and line of business applications will generally be non-VSS capable and in this case SCDPM will:
On a VSS capable guest operating system.
This referential VSS writer process means that:
Virtual Server exposes the backup options for VSS and non-VSS capable as online or offline backup. Hyper-V is more descriptive, with Backup using Child Partition snapshot (the equivalent of an online backup) or Backup Using Saved State if there are no integration components available
So, with no downtime and no agent deployment for each guest operating system, why wouldn’t we always protect virtual machines from the host? Well, when we protect the guest from the host, the whole virtual machine is treated as a logical unit without any data selectability or granularity. Whilst there are some advantages to this approach (it allows bare metal recovery of virtual machines to any other host; the whole virtual machine set can be protected with a single SCDPM agent and a single DPML license; non-Window or legacy Windows operating systems can be backed up), if an agent is deployed within the guest then SCDPM can select the data to protect/recover – e.g. individual SQL databases, Exchange storage groups, file sets, Sharepoint farms, etc.) but with the additional cost of deploying and licensing agents.
We can also use a hybrid of the two models – running an agent inside critical virtual machines but only using host-based backups for non-VSS cpabable operating systems. Indeed, it may even be desirable to protect the entire virtual machine and its data separately (e.g. if using passthrough disks then the guest operating system cab be backed up via the host and the data protected via the guest).
SCDPM 2007 SP1 will back up shares, volumes, system state and any VSS-aware application workloads (subject to licensing options – an Enterprise Data Protection Management License will be required for native backup and recovery of applications). On the licensing front, it’s also worth considering the Server Management Suite Enterprise (SMSE) as it includes management licenses for System Center Data Protection Manager, Operations Manager, Configuration Manager and Virtual Machine Manager with free usage rights up to the number of guests licensed with each host operating system edition (one for Windows Server 2008 Standard Edition, four for Windows Server 2008 Enterprise Edition and unlimited for Windows Server 2008 Datacenter Edition).
The key points for agent placement are application consistency and the granularity of restoration required. By deploying an agent inside a virtual machine, a VSS-aware application will be signalled that a snapshot is about to be taken and consistency can be guaranteed. Alternatively, if application consistency is not an issue, by installing the agent on the host, each virtual machine can be backed up as a single container – in effect the virtual machine will be consistent but the application may not be.
This afternoon, I finally got around to rebuilding my everyday notebook PC (a Fujitsu Siemens S7210) onto the Windows 7 release candidate (build 7100, 64-bit).
What follows are just a few notes that I made along the way:
bcdedit.exe.diskpart.exe) but that may be a limiting factor later.Tomorrow I’ll be in the office, so I’ll join the domain and start working using a normal (unprivileged) domain user account. The most important point from this exercise was that everything I needed installed without issue (drivers and applications) – even though this is a 64-bit installation. It’s inevitable that there will be some legacy applications that don’t work though – even though I have a pretty modern application stack, there are some legacy applications in my corporate infrastructure that might need me to revert to IE6, for example (using Windows 7 XP mode) and I’m also working out the best VPN option, as the Cisco VPN client is 32-bit only. Over the next few weeks, I’ll see how things progress but my initial impressions of the Windows 7 release candidate are extremely positive. This feels more polished than the beta – and way better than Vista. I’m looking forward to working with my customers to help them migrate from legacy XP desktops that really are no longer fit for a modern computing environment.