The Knight Capital event and a DevOps Discussion

August 15, 2012 Paul Peissner

I have been intrigued by the recent Knight Capital stories, articles and commentaries especially by members of the DevOps community (see links below).

Here is my quick summary on the Knight Capital story…

The company, Knight Capital, is in the fast pace and highly competitive trading market.

– They run complex automated trading systems, trading services (for options, currencies and bonds), and manage order execution for 19,000 equities.

The IT (and software) environment at Knight Capital is one which has…

– a fast paced, highly visible, complex environment that require regular upgrades

– a mature “change process” governed by regulatory guidelines like ISACA Cobit and established their own best practices

– accountability to Securities and Exchange Commission regulators

– potential far-reaching impact when incidents occur

– large impact when even a small outage or short downtime (caused by a normal “glitch”) occurs…

1) Long-term impacts on their corporate reputation in industry

2) Immediate revenue losses (for company and customers)

3) Global attention and inedible scrutiny brought down on their technologies, teams and processes

The event…on August 1, 2012, was a chain reaction to an upgrade incident that…

– resulted in documented losses for the company

– created undesirable “automated” trades with their systems, costing $440M

– exposed a “system / process” flaw created a series of bad trades

– high-lighted a need for improved best-practices and organization-wide collaboration and

– initiated a series of investigations (internal and external) on how and why

My take…

Today in many industries the intimate relationship between software and business success in competitive markets is inseparable. For many of us software innovations from previous generations, legacy processes, historic change cadences, outdated best-practices and old-school collaboration standards are at the end of their life-cycle. The lack of Software and IT adaptability is driving up IT costs, polarizing teams, creating technical debt, slowing down business agility, hurting reputations (internal and external), decreasing innovation and eventually creating situations where quality is compromised and services disrupted.

So where do we go from here? How do we evolve, transform, transition, modernize, change and support the revolutions that will help our business organizations pursue the highest standards of competitive agility? Is DevOps “the” answer?

Clearly, companies like Knight Capital continue to adapt and integrate in new IT approaches. And DevOps is helping those interested in surviving and thriving in this period of dramatic change. But I also hear some people talk about DevOps like there’s a Silver Bullet.

DevOps is not…

– DevOps is not a technology, a one-size fits-all tool-chain or a vendor platform to buy or install.

– DevOps is not another specialized silo team or snapped-on department function.

– DevOps is not a specific job title/role or secret knowledge you need to “own” or hire.

– DevOps is not a single process philosophy that needs to be taught and religiously adhered to.

– DevOps is not part of the innovative DNA found only in the new generation of (in-experienced) leaders.

DevOps is…

DevOps is more of an internal IT, or Software conversation every organization should explore (people, processes and technologies) to ensure there is no break-down in services or missed opportunities for the business. An open-ended DevOps assessment conversation should help an organization…

1) Question presuppositions of current (normal) culture, processes, structures, technologies and decision-making criteria

2) Look for blind-spots in established best-practices, organizational assumptions, and organically created processes

3) Explore and discuss changes and innovations across the industry that would benefit the organization

4) Facilitate conversations and knowledge sharing around what can or cannot be changed (enterprise architectures, security, compliance…) and why.

5) Establish shared-goals and unified efforts to…

  • Optimize your business’ competitive agility, profitability and long-term success
  • Maintain the highest standards of quality, security and performance
  • Improve culture, collaboration, more holistic production processes (including manufacturing, deployment, maintenance, quick fix, etc.), encourage any innovative advantages, review decision-making processes and courageously explore risk-appropriate business-benefiting changes (i.e. automation, re-use policies, cloud adoption, mobility…).

I would invite you to bring your knowledge, experience, concerns and insights to explore the industry-changing DevOps journey and discussion. Otherwise, you might be driven directly, or indirectly, to this conversation by an inevitable chaotic event, born from the ongoing clash between loyalty to legacy innovations and investments and the continuous evolution of modern innovations.

Do I believe the Knight Capital event could have been avoid with an internal DevOps IT initiative? Absolutely, but it would have taken a courageous and visionary team, not simply interested in maintaining a “status-quo” legacy IT practice.

Recommended Knight Capital Article Links:

  1. Chicago Tribune article
  2. NY Times article former developer opinion perspective
  3. Sticky Minds DevOps commentary by Bob Aiello
  4. HP Blog DevOps commentary by Rafal Los

Previous Article
Git, Subversion and Android Development for Embedded Systems
Git, Subversion and Android Development for Embedded Systems

NXP is a leading provider of embedded systems. While NXP is relying on Subversion as the corporate standard...

Next Article
What Does Done Mean
What Does Done Mean

We all know that we need to create business value in each iteration. But what does it mean to be “done”? Wh...