Heard on the Street: IT Complexity Has Run Amok
There was something heartening about the standing-room-only crowd at HPC on Wall Street, the annual event which was held this week in New York City. I didn’t personally attend last year’s event—which coincided with a global financial meltdown—but I can imagine the mood was sanguine or worse. This year, however, the mood was decidedly upbeat. Very little gloom and doom. In fact, the tone was almost celebratory, as if to say: There is a future to behold!
What a difference a year can make.
If the nominal theme of last year’s event was “survival in the face of chaos,” this year’s event may have been called “thriving in the face of complexity.”
The word on the street last year was that the dire predictions may be understated.
The word on the street this year was that we’re back, but complexity has run amok.
The event kicked off with keynote addresses from Madge Meyer, EVP of global infrastructure for State Street Bank; followed by Erik Troan, founder and CTO of rPath. In the packed ballroom of the Roosevelt Hotel, both presenters discussed complexity in large-scale IT environments.
Madge’s presentation was part IT strategy and part leadership seminar, a thoughtful message about “doing more with less,” which is fundamental to Madge’s role—and familiar to any other IT leader today. How do you take on escalating demand, at progressively higher service levels, with declining people resources and diminishing IT budgets? Oh, while also remaining compliant?
The picture Madge painted is the backdrop of any large IT organization today.
The problem, of course, is that IT is becoming unthinkably complex. The combination of system scale, software diversity and the accelerating pace of change are forcing IT to the brink.
Erik took this on directly during his keynote, suggesting that the current state of systems management will not scale to meet the needs of HPC and other large-scale enterprise environments. He suggests that the problem lies in today’s manual approach to creating, provisioning and maintaining systems, which is severely pressured by three trends:
- System instances are proliferating – HPC and other enterprise scale-out environments mean more system instances to manage, which puts pressure on the traditional cost structure of systems management. Throwing bodies at the problem is no longer an option and virtualization and cloud promises to make it worse by compounding the number of systems.
- Software is becoming increasingly diverse – Between open source components, frameworks and dynamic programming languages, software developers have extraordinary resources for plying their craft. This is making their jobs easier, but all of the variability is making it harder than ever for IT to deploy and maintain applications.
- The rate of change is accelerating – The adoption of Agile and other iterative development methods mean more rapid and frequent change. IT needs to find a way to consume this change, while also facing the reality that—today—change is exceptionally disruptive to IT.
And it’s this inability to consume change that is making complexity such a risk for IT operations today. According to Troan, “App dev has made change part of their process and IT still treats it as an exception case. For IT, change still means pain. They need to find a way to make change a business enabler rather than a business risk.”
This means rethinking systems management based on three new rules for tackling complexity:
Rule #1:Application Centricity – Shift the orientation from infrastructure to applications.
IT spends too much time focusing on infrastructure when only applications deliver business value. Determine the infrastructure that applications require based on deep dependency analysis and use this understanding to limit the scope of infrastructure you deploy.
Rule #2: Version Control Everything – Manage the entire system definition under strong version control. This provides a basis for understanding the impact of change, easily reproducing exact systems over time and rolling them back when applications break.
Rule #3: System Automation – The combination of fully described applications and system version control provides the foundation for complete system automation. Now, system creation, provisioning and maintenance can be fully automated, providing a scalable solution for escalating complexity.
Erik’s view is that automation is the perfect counterpoint to complexity. But it requires a deep understanding of systems—a model—and a version control foundation. Without this, IT is blindly automating systems they don’t fully understand—and without a means for reversal.
If 2009 was the year of survival, 2010 will be the year of simplification. The financial crisis reset the rules and forced us to look differently at IT cost and complexity. Going forward, new rules must be applied to address the changing economics of IT and complexity run amok.
Comments
Leave a Reply
