Despite the continued success of cloud-based business operations and the growing trend of migration to the cloud, many companies remain skeptical and even untrusting of cloud-based operations, and of Office 365 in particular.
Microsoft only has itself to blame for this hesitation, as the past five years of its Azure cloud's history includes a litany of underperformance complaints, app crashes, me-too feature rollouts and customer frustration over the opacity of availability failures. Azure isn't a bad cloud environment by any means; in fact, it has become a pretty good cloud environment. It just isn't yet great -- or is it?
This year's Ignite conference was Microsoft's opportunity to demonstrate that recent improvements to Azure, and Office 365 in particular, meaningfully address customer concerns about Office 365's back-end accessibility and Azure's need for transparency.
The core problem -- that customers become uncomfortable when business processes are interrupted or are not performing as expected, and that they do not have access to the servers running them -- is what Microsoft needed to address.
Change monitoring and configuration services. A huge issue with trusting software and storage to vendor servers is having to cross your fingers when a vendor introduces patches and upgrades in their own environment. Will those patches and upgrades sink any client apps? This can happen (too frequently) in SharePoint-hosted custom code, even if that code conforms to the now ubiquitous app model standard.
The Microsoft Operations Management Suite now features services that address this server blindness and surprise factor. Changes in the server environment can now be observed by the customer, which is tremendously useful, even though the customer still can't touch the servers themselves. With the Operations Management Suite, customers can more easily foresee outages and anticipate degraded service. Moreover, root cause analysis -- until now, a leading barrier for customers to deploy apps in O365 -- is finally being accommodated, significantly reducing time to resolution when apps fail or underperform.
Log analytics and automation. Because cloud users can't see or touch the servers on which they're running their operations, they don't have access to the activity logs that are so invaluable for troubleshooting apps and diagnosing issues. This has been a common complaint of Office 365 users, and a barrier for many who might otherwise adopt it.
Whatever is said in public, the reality is that no cloud provider really wants customers nosing around in logs that might betray an actionable error or any malfeasance on the part of the provider.
Microsoft's middle ground here is customer access to searchable, real-time and historical machine data in Azure. This enables customers to troubleshoot app crashes, diagnose performance issues and set up automated responses to real-time problems.
The catch here is that Operations Management Suite is designed for non-specialists, with canned code and an obscure schema. It's reasonable to predict that its effectiveness will be somewhat limited -- but even so, it's a big step forward.
Also, while living in Azure, this product also works in VMware, Windows Server and other OS environments.
Detecting system and service issues. BlueStripe has been renamed the Application Dependency Monitor in the Operations Management Suite. In an effort to get their arms around the transition of the service from a custom code 'sandbox' to a new (more efficient) app model, Microsoft has redesigned it to centralize connections between customer apps, services and data sources, even between clouds.
If the Application Development Monitor works as advertised, it will be a major relief for those wishing to deploy custom functionality within O365 (and Azure in general), even in the case of complex business processes that touch many different servers and sources.
As a diagnostic aid, this is a significant step forward: Understanding and monitoring traffic between system components in an environment that is essentially server blind to the user was beyond reach until now. Mapping dependencies, let alone making them monitor-friendly, could make troubleshooting apps, processes and systems far more doable in the cloud than it has ever been before.
Application disaster recovery. Finally, a major issue with trusting the cloud for app deployment is the customer's helplessness when systems crash. If they can't touch the servers, they can't control the recovery. This gets even more complicated in the hybrid cloud world, where apps/processes are integrating ground-to-cloud or cloud-to-cloud.
The Operations Management Suite now features dedicated Azure Site Recovery, which simplifies recovery planning, time and expense. In the case of fully customized apps and processes, this is an ideal solution, and makes recovery faster and cheaper than is usually possible with on-premises solutions. Moreover, this is an inexpensive and truly viable dev and test alternative.
In the case of Office 365 app deployments, it's not so clear cut. Office 365 apps live in a more controlled environment where many customers share the same servers; disaster recovery, when it's software-specific, is aided by the new functionality detailed above -- but physical server issues impact more than one customer. As a result, recovery is more obscure from the user's standpoint.
Even so, this is real progress. The announcement of the Operations Management Suite at Ignite signals that Microsoft is listening to existing and potential Office 365/Azure customers. This gives customers good reason to expect meaningful change in the cloud leading to growing user confidence.
Power BI with SQL Server
Navigating the Microsoft roadmap
Pros and cons of Power BI
Discover more about mobility Office 365 suite