Original URL: https://www.theregister.com/2014/06/22/microsoft_cloud/
How Microsoft's cloud aims to cover the world
A three-pronged strategy
Posted in SaaS, 22nd June 2014 23:59 GMT
Microsoft has a vision for the cloud. Anyone who has watched this session by the legendary Rick Claus from last year’s AUTechEd already has a basic idea of what this entails.
Microsoft’s plan, dubbed the Cloud OS platform, rests on three pillars. These create a unified strategy that is changing the game and positioning the platform as the most complete cloud service available.
Define your terms
Let’s establish right now exactly what I mean when I say cloud. I am the son of a meteorologist so when someone says cloud, the first word that comes to my mind is "agglomeration". The definition of an agglomeration is “a heap or cluster of usually disparate elements”.
In mathematics, a cloud is “a large number of points in a coordinate system”. So if we combine the meteorological definition with the mathematical definition we get this: “Cloud computing is a computing architecture made up of many individual yet fully coordinated elements, which appear to the casual observer to be a single object.”
Got it? Fantastic. Let’s get back to what you came here to see.
Private cloud
When most people hear private cloud, they think “Oh well, that’s just a funky new name for what we are doing on premises already.” If you are one of those people you would be partly right but mostly wrong.
In its strictest terms, you have a private cloud only if none of the data goes outside your corporate firewall. A private cloud is often created using existing hardware through a combination of a cloud-aware operating system and a cloud-aware management tool.
In the case of Microsoft this refers to servers running Server 2012 or 2012r2 and being managed by System Center 2012 or 2012r2. It is not enough, though, simply to run an operating system.
If you are in a single-server environment running Server 2012r2, you don’t have a cloud. Conversely, you might be managing a fleet of physical servers (each running a single operating-system install, with no virtualisation at all) and you still don’t have a private cloud.
So what makes a private cloud different from virtualisation? If I had to sum it up in one word it would be orchestration.
Orchestration is the combination of a management console, a managed or self-service portal and virtual machine templates. When combined, these three elements allow your clients (whether employees, contractors or external clients) to create, or “spin up”, their own virtual machines.
Depending on the hardware you are running and the requirements for the virtual machine, a template-created virtual machine could be online in as little as three minutes.
At the basic level, what private clouds offer above all is control. If what you want is complete control over all of your infrastructure, you have that because you own it. If you need complete control over your software, well you have that too because you install and configure and manage it.
You could, if you wanted to, walk up and hug your servers. We don’t recommend this, but you could.
Public cloud
Public cloud, as it refers to Microsoft, is a combination of Office 365 and Azure.
When we think Office 365 what most of us think about is the end of Office as a standalone product, but it is far more than simply a new way to pay for a productivity suite.
Office 365 is at its core all about speed. Think how long it used to take to roll out a brand new Exchange Server.
With Office 365 I can do that in about half an hour. Thirty minutes and I don’t have to worry about the back-end infrastructure, managing database availability groups or what happens if someone disables IPv6 on my Exchange server.
Where do I sign? In the bad old days to add capacity or roll out a test environment for a new version of Exchange Server might have meant taking staff away from business-critical work for days, if not weeks. Not any more.
Azure was born out of that same need for speed. I work in a vertical with especially stringent federal statutes. We have 40 clients who run the gamut from small two-workstation one-server installations all the way up to 80 seats and multiple servers.
To have a test lab running at all times that caters to every single site would require a large number of servers, and rolling out a test environment could take hours. With Azure I can spin up an environment in minutes, do the testing work that I need to do and then destroy that environment.
Azure even makes it easy for me to run virtual machines and the virtual workloads they entail inside that testing environment.
With Server 2012/2012r2, the Azure integration goes even deeper. If I have configured it correctly, I can use Azure to cover those times when a non-cloud integrated infrastructure would be brought to its knees.
By bursting out to the cloud I can use Azure as my cold-storage failover site for disaster recovery. I could use it to handle expected (or even unexpected) periods of intense bandwidth.
Who will provide the service?
Service provider is the last of the three pillars but by no means the least important. For years Microsoft has supported a vibrant channel and built an ecosystem that encourages its partners to make money by easing access to Microsoft’s products.
As the way we access our networks changes, so too must the way those services are offered to us. This is where a service provider steps in.
At this point I want to introduce you to a new term. We have all grown accustomed to calling anyone providing managed services an MSP, but not everyone who offers cloud services is an MSP. Enter the cloud service provider (CSP).
CSPs are the new black, but many are also the traditional MSPs that you have been using for some time, just rebranded and with a newer, cloudier industry term.
CSPs are the guys Microsoft is betting on to provide vertical integration and localisation of its cloud
CSPs are the guys Microsoft is betting on to step up and provide vertical integration and localisation of its cloud.
In Australia, for example, these are the companies that will customise a solution to fit the healthcare industry where compliance with federal regulation is difficult and expensive.
Microsoft is clearly happy to provide the platform and let each CSP customise it to meet the needs of its clientele.
This has an added benefit. To offer a product that was compliant with every single regulation in every single statute in every single jurisdiction that it operates in would be beyond the reach of Microsoft, even with the massive resources it has at its disposal.
By allowing CSPs to customise and individualise the platform, it gains access to markets that alone it would not be able to cater for, and still keep the mindshare.
This is where we talk about the other reason you pick a CSP. A lot of sysadmins and companies that I talk to about cloud services say the same thing: “We would use cloud services more for our backups, but how do we then get our data back in a hurry?”
This is where CSPs can really shine. As a friendly sysadmin from Canada points out, while it may not have hit our shores yet, local CSPs in his country actively advertise that they offer two major advantages over the multinational competition: no US attack surface (great from a liability standpoint); and the ability to drive to the CSP’s office in the nearest major city and collect a copy of your data on a HDD when the manure hits the fan.
At their most basic levels, service providers offer an easy way to keep your data local and ensure that you (and they) are compliant with all relevant law. This is a big deal as compliance issues affect everything from data sovereignty, localisation, data retention and in some rare cases minimum requirements for disaster recovery procedures.
Hybrid solutions
When most vendors talk about hybrid cloud they talk about it as an all-or-nothing option.
You either pick all-in private cloud, all-in public cloud, or a static level of both and call it hybrid cloud. It can also be difficult, extremely time consuming and in some cases nearly impossible to alter the state or level of your dependence on private and public cloud once you have implemented it.
These static levels are really a throwback to a time when office workers had a single workstation through which they accessed everything necessary for their daily tasks, but clearly it is not a situation that exists now. Let’s look at an example.
Client A is a growing business. In the past year it has grown from five sites and 40 seats to seven sites and 65 seats, and it is planning to add another 20 seats very soon. We have just migrated this client to a Windows Server 2012r2-based virtual environment because it had exceeded the limitations of the XenServer and hardware it was running on.
Within the next three months we will move it to a much bigger, more robust environment. When that happens, this company is a prime candidate for a hybrid cloud solution involving us, as a CSP, offering it a cloud-based backup solution.
We have knowledge of its vertical and can build a solution to the issues created by several of the more dastardly statutes it needs to prove it is compliant with.
We can also do this because Microsoft decided not just to build a single component or pillar, but to go the whole way and build a platform. Its One Cloud OS vision has allowed us to offer services to our clients that other multinational cloud providers could never hope to offer.
They don’t know our vertical, they don’t know the laws and statutes in place and they don’t have the years of boots-on-the-ground experience required to build the custom solution we supply to these clients.
One Cloud for all
The One Cloud OS platform means that everything Microsoft does is inter-linked. Everything that Microsoft learns from its Azure team feeds through to the Windows Server team.
All of the feedback from its service providers, whether about the management tools or the interface or the stability of the product, is incorporated into all of the products.
This is a radical departure from how the Microsoft of old used to operate and it is a welcome and long overdue change.
By using a single shared code base for all three offerings, Microsoft has at last moved to a virtuous cycle of development. This has enabled the company to iterate faster, implement patches and fixes more efficiently and to offer new features and improve existing ones far faster than before.
Windows Server 2012 is now on a yearly release cycle. Azure gets new features monthly and a major release approximately every quarter.
Microsoft has managed to reach the holy grail of the cloud space: it has built a single consistent platform that crosses all previously established borders.
The One Cloud OS is a platform where integration between private cloud, public cloud and service provider cloud is built in from the ground up.
More than anything else though, it has leapt ahead of the competition by building a consistent platform of products that make it easier for those of us on the ground to do our jobs. ®