"The Datacenter" is moving to the hybrid-cloud. So much so that it might as well be called the Hybrid Data Center. While the hype around the growth in the Public Cloud continues to grow, Amazon's moves to link AWS with enterprise applications and the growth of Microsoft's Azure, and Microsoft's introduction of Windows Azure Pack both highlight how the "new" enterprise environment will be a hybrid – with on-premise physical systems, private clouds, and public cloud services all contributing components to application architectures.
For IT organizations, especially in the enterprise, the result can be unparalleled flexibility and the ability to scale capacity – that's the vision. Taking advantage of this flexibility, however, will require application-level visibility that reaches across all the different platforms and operating systems to deliver a single view of application performance.
It has been said that every company in business today is a technology company (http://blogs.gartner.com/mark_raskino/2013/11/28/every-company-is-a-technology-company-more-and-more-evidence/). Even companies in the most "old-economy" industries have websites, ERP systems, payroll, etc., and more are integrating technology deeply into their products and services. We're all tech entrepreneurs now.
At the heart of the move toward technology ubiquity is the distributed application. Distributed applications – applications that use many servers to accomplish their purpose – are what make large scale websites and web-enabled services possible. The size can range from a simple 3-tier webserver/app server/database style website to massive scale services that combine thousands of servers with petabyte sized databases.
Given that distributed applications are critical to business success, it makes sense that organizations would be reluctant to risk pushing their components out to the Public Cloud. Critical infrastructure needs to be under control. The bulk of Public Cloud growth has come from development/test environments, short-term batch workloads, and only occasionally a large-scale net-new enterprise service. As confidence in Cloud grows, organizations are taking commodity functions and placing them in the Cloud, while maintaining control of other functions with on-premise servers. Thus, we now have Distributed Applications running in Hybrid environments. Voila – the Hybrid Data Center.
Unfortunately, managing distributed applications is complicated. There are many components involved – the actual application code itself, plus several different flavors of infrastructure (web servers, application servers, databases, middleware – all running on different operating systems), IT services like DNS and Active Directory, and of course the network that ties it all together. Software vendors have been happy to supply tools for managing each of these components individually, but tools that provide the big picture view of the entirety of the application and its performance – across the entire infrastructure – are rare.
Most organizations get around this through brute force – bringing together the subject matter experts for each technology whenever there is a serious performance issue. Visibility into the entire application structure is limited at best. For many organizations, the prospect of adding a Cloud-based component to the mixture requires will only further complicate an already murky picture – unless they can achieve better application-level visibility.
At BlueStripe, we work closely with a wide range of enterprise IT teams. A key takeaway from these administrators and executives is that application visibility is critical. These organizations cannot simply manage server performance and availability – to thrive in the Hybrid Data Center, they must adjust to complete application infrastructure and performance visibility. In fact, from Development to QA, Product to Support, and IT to the Business, forward-looking organizations should have complete transparency within their service delivery process.
The next question is "what exactly does end-to-end application visibility look like in a Hybrid Data Center?" It's actually quite simple:
- See hop-by-hop performance across all servers on-premise in the Data Center
- See hop-by-hop performance across all servers hosted in the public Cloud
- Track the cross-over points between the Data Center and the Cloud
And, of course, tie them all together into a single view from the user request through all back-end systems and back – application by application.
A primary benefit of end-to-end performance visibility for distributed applications is that it gives technology owners (for each tier) the ability understand how their systems work together to deliver business services. It allows QA and Production Support to discuss issues with Development teams from a clear and agreed on perspective of how systems are working together.
And it allows IT and Business executives to understand whether or not the organization is delivering the appropriate business services to the target number of users at the appropriate speeds.
Hybrid Cloud is like the proverbial cat that's out of the bag. Once implemented for a business application, it's just not going to go away. IT leaders, especially Operations executives, MUST adjust their thinking about how to manage business services, or else their teams will spend the majority of their time managing individual resources without a notion of whether or not the systems, themselves, are actually meeting their goals.