But the big news from my perspective is that by the end of 2012, two of Microsoft's ERP products, GP and NAV, will be available on Microsoft's Azure cloud.
Click on the video interview at the right for my initial thoughts, which I am expanding upon in this post.
Azure Complements Existing Hosted OfferingsMicrosoft customers have always been able to deploy NAV (formerly, Navision) and GP (formerly, Great Plains) on-premises. In addition, some customers have chosen in the past to have Microsoft partners host their systems in partner data centers. MyGPCloud is one of the largest such partners, hosting GP for thousands of small business customers. Likewise, Tribridge offers similar hosting services for all Dynamics ERP products.
Now, Microsoft is offering customers the option to deploy their GP or NAV systems on Microsoft's Azure cloud, which runs in Microsoft data centers. This offering will not replace partner hosting but simply will be another deployment option for customers.
Through back channels, I've heard some partners express uncertainty about this new development. Is Microsoft attempting to go direct with customers? How will the partners make money? During the session, Microsoft executives made clear that, under Azure deployment, partners will still maintain the customer relationship and deliver the services for implementation and ongoing support. The only difference is that with the Azure deployment option partners will be relieved from the need to maintain data center infrastructure.
What Are the Benefits?Over the past two years, I've been one of those encouraging the Dynamics team to go faster in moving to Azure, as cloud ERP is already available from competitors. But now that Microsoft is on the verge of actually doing it, I wanted to know, what are the benefits? Specifically, if customers can already have these systems hosted by a Microsoft partner--and if Microsoft will still work through partners in selling and supporting systems deployed on Azure--what are the added benefits of Azure?
I asked this question a year ago at Convergence and, frankly, the answers were not that clear. After asking this same question in several briefings this year, and adding my own analysis, I think the benefits picture is now emerging.
- Azure deployment is cheaper than hosting. Azure is a true elastic cloud platform, with data center economies of scale that traditional hosting cannot come close to matching. This should allow Microsoft to price these services at a lower cost than what partners can offer.
- Azure deployment scales beyond partner hosting. As a true cloud platform, Azure deployments can scale instantly beyond what partner hosting can offer. Hosted ERP relies upon dedicated resources, which must be planned and expanded manually to meet changing customer requirements. With Azure, customers will never exhaust the resources available.
- Azure supports worldwide deployments better than partner hosting does. Microsoft runs Azure data centers worldwide and can move customer systems and data between them as needed. Hosting partners do not have this capability, unless they are utilizing a true cloud IaaS, such as Amazon's EC2. The move to Azure is therefore a better choice for organizations that are running separate instances in different parts of the world.
- Azure deployment provides easier version upgrades. With partner hosting, upgrades and maintenance are handled more or less as they are with on-premises software: each customer is treated separately (though I suspect some partners are more organized about this than others). With Azure deployment, Microsoft will have a more disciplined approach to application management: rolling out new versions, upgrades, and patches to its customers, similar to what it does today with Microsoft CRM (even though, as I point out in the interview, CRM is not yet an Azure service).
- Azure deployment is provided directly by Microsoft. Most new prospects will have a higher level of comfort with cloud services provided directly by Microsoft and backed by the Microsoft brand and service level guarantees. Hosting is often delivered by service providers who are relatively unknown. The direct Microsoft relationship is also simpler and easier to explain. The software comes from Microsoft and the cloud services are delivered directly by Microsoft.
Optimizing Azure as a Cloud PlatformI am struck by the fact that I've had to piece together this value proposition for Azure ERP myself, lobbing softball questions to Microsoft executives, parsing their answers, and adding my own analysis. If Microsoft itself is not prepared to articulate the value proposition of Azure ERP, how can it expect that its customers or its partners will perceive it?
Therefore, I do not envision customers and prospects staging a mad rush to Azure. As I said in the interview linked above, what if Dynamics throws an party and no one shows up?
Nevertheless, from a strategic perspective, I do believe that moving to Azure is the right thing for Dynamics. Mike Ehrenberg, one of only a handful of Microsoft Technical Fellows, told us an interesting story. He said that when they first spoke with CEO Steve Ballmer about moving Dynamics ERP to the cloud they told him that they could do it in one of two ways:
- The quick way: hosting it in Microsoft data centers in a highly virtualized environment, as they had done with Microsoft CRM, or
- The strategic way: working with the Azure team to optimize the Azure capabilities needed to support true scalable enterprise business applications, such as SQL Azure, until it could support Dynamics ERP.
The problem, of course, is that it's taking much longer to develop Azure as a enterprise-class platform. In the meantime, competitors such as NetSuite, Workday, SAP, Plex, and others have already become established as cloud ERP providers and have gained market share in this emerging market. Nevertheless, Microsoft entering this market later this year is a welcome development that will mean an increasing number of choices for buyers.
Postscript: watch for Part 1 of my market overview of cloud ERP over the next few weeks.