The Internet is a big place and the demands on services can be tiny or massive and this is expected to change over time, sometimes within minutes. A cloud solution should have that flexibility built into it. When we deliver computing resources in our own offices or data center, we have to buy the right number of servers and storage to enable us to meet the peak demand. Probably, because it is too complex to do anything else, most organizations simply leave these servers on all the time, consuming electricity, but adding no value to the business. In a cloud environment, servers are started and stopped as needed to deliver the solution, often by the service provider on our behalf. We can see all the opportunities that an elastic service can simply scale up or be turned off according to our business needs in the following diagram:

In the case of Windows Intune, this is the application that Microsoft manages and they start up and stop the servers as required to deliver the information to the Windows Intune client software on the PCs and the management interfaces.

There are two elements to the pricing of cloud solutions that need to be discussed. The first is that by virtue of the elastic number of servers used, the cost of the service is lower than purchasing for peak capacity. The second is that rather than paying for servers, power, cooling, storage, and people, we now pay for utility. This may be a fee per user, compute hour, storage, and so on, but it is nothing like the traditional buy or lease models for IT where we paid for the hardware and software rather than usage.

The low fee we pay for Windows Intune relates to the fact that the service is delivered at Internet scale with elastic scalability that means that the actual computing required is right sized for all the users of the service at any moment. Our usage would be a fraction of the IT required to deliver ourselves because we do not actively interact with the management system 100% of the time. Beyond the scalability, Microsoft is also able to deliver greater efficiencies in managing the whole solution. This enables Microsoft to charge a flat rate per user for the services that is lower than the cost of acquiring the software, hardware, and people to manage by a significant amount.

Protocols describe the way machines talk to each other. Some of these are defined by standards bodies and others are known as de-facto standards as they have been popularized by the organization or organizations that use them. Either way, they need to be Internet friendly so that they can be routed, inspected and secured by firewalls and received by different client device types where appropriate.

Windows Intune uses a combination of standards based and de-facto standards, but all the protocols are published and known, enabling routing and securing via the Internet. Windows Intune uses a standard web browser to deliver the management interfaces and then a well-known set of protocols to deliver the rest of the information to the client PCs.

The term Cloud is often used when describing solutions that are often hosted and certainly connected to over the Internet. What used to be an Internet service is now often referred to as a Cloud service, but the terms are often interchangeable.

The final technical facet that all cloud services should have relates to security and privacy. The service needs to be as trustworthy as if we were running it ourselves. This means that our data is not shared or leaked, and that the communication protocols are secure. This is often not something we can test, but we should look for a privacy statement from the service provider.

Windows Intune's privacy statement can be found by following the link: http://davidoverton.com/r.ashx?20.

All quality cloud services should have a Service Level Agreement or SLA for short. This will need some careful analysis as this is one of the areas that differs the most from running the technology and delivering the service ourselves, and having it provided as a cloud service.

Service Level Agreements consist of two elements repeated for each different feature or function of the service being offered. These two elements are Availability level and Service hours of a time period and, as a customer, we get some form of Recompense if the level of service is not delivered. As we are managing multiple users, the number of users impacted, or the proportion of them that were impacted is often also included in the calculation.

Window Intune SLA

We have covered the Windows Intune SLA several times in the preceding text and there is a snippet of this document below. To get the latest version of this, please go to the following URL: http://davidoverton.com/r.ashx?21.

In this snippet, we can see how Microsoft calculates uptime at the time of writing. We can see that Microsoft considers the number of users impacted as well as the number of minutes that they were disrupted to be key to calculating the Availability or"Monthly Uptime Percentage".

We can see that if the service impacts all users for more than 44 minutes a month, we are entitled to a 25% service credit. In my opinion, this is one of the best value service credits available in the marketplace today.

We are now moving on to requirements that are not always present in all cloud computing scenarios, but it is important to understand how they might be useful.

Not that many years ago, the de facto desirable device for all computing needs was a Windows based PC. While the benefits of Windows have increased, as has the use of Windows 7, people want to be able to access information and applications from other devices. We have seen the rise of netbooks, tablets and phones as information access devices.

These devices all have relatively small storage and processing capacity, so the natural thing is to store information and provide the computing power on the Internet where it can be delivered on mass as needed. The services that provide all of this are often delivered using cloud computing solutions as they need the facets discussed in the preceding text.

Once the key information, data, and processing are deployed in the cloud, all these devices can utilize it from a browser or a small application that calls upon these cloud services. This means that rather than having to write complex software for all these devices, which has always been a barrier to deploying on anything but Windows in the past, now it is developed once for the cloud environment and delivered as a simple piece of consumption software for each device type. As the consumption software is lower in complexity because the complex solution is now being delivered by the cloud service, it becomes cost effective to write it multiple times, once for each device.

Now we have a model that allows access from multiple devices from any location provided they have a possible connection to the network with the cloud service on it, normally the Internet.

iPhones, iPads, Android, Windows Phone 7, netbooks, notebooks, Windows XP, Vista, and Windows 7 PCs are now all excellent consumers of cloud services. Each device has its own interface quirks, tools, and other management issues, but for a cloud service provider, they can now address more devices and users than ever before.

Today, Windows Intune is only used to manage Windows PC devices and the management interface is only available through a web browser with Silverlight installed, however, I expect more interfaces to become available over time.

Cloud computing requires a way for us, the user, to interact with it. This may be via a web browser, but if that is not the case, then we need to ensure the device we are planning on using has the software available to download in a marketplace, or application store for the specific device. This could be an application on a phone or computer. Familiar examples would include an e-mail client or music playing software.

In the case of Windows Intune, we need client software on each Windows device that will be managed. The client software communicates to the cloud service and we manage the service via a web browser.

With a cloud computing solution we have much of the data and clever computer processing being done in a data center across a network, which means this now has to be controlled remotely. There are two choices here as to how this is done, either by building the whole service and infrastructure and solution as the vendor, or by enabling third parties to have access to the control interfaces.

For some solutions, having third parties access these control interfaces makes perfect sense as it enables additional benefits to users without having to spend the time building it ourselves as the vendor. For some solutions however, the data is considered too complex to manage or too important to let third parties change, so the interfaces are not accessible.

Windows Intune today does not have publicly accessible programmable interfaces, so no enhancement of the software is possible

If these interfaces are available then the vendor will also have some form of additional shop front or application store to enable us as the user to choose to augment the cloud computing solution we have.