February 2015 Summaries

Balancing delivery and quality in software development is a persistent challenge, with developers often spending around 75% of their time debugging and fixing production issues. While common solutions such as unit testing, code reviews, and adhering to coding conventions are well-known, the article suggests that leveraging log analytics can significantly reduce the time spent on these tasks. Quality logging involves creating informative log entries that narrate the system's activities and using metadata to categorize them, which aids in efficient problem-solving. Given the vast amounts of log data generated daily, log analytics companies are emerging to assist in making this data searchable and providing tools like alerts and trend graphs. These companies aim to enhance their offerings by automatically identifying problem causes and potentially providing actionable solutions, a trend expected to grow with advancements in Big Data analytics.

In the software development industry, a significant portion of developers' time and company budgets is spent on debugging and fixing code, with statistics showing that 75% of a developer's time is dedicated to these tasks, equating to approximately 1,500 hours annually. This focus on debugging arises from the prevalence of errors, with an average of 70 bugs occurring per 1,000 lines of code and 15 of these reaching customers. The process of fixing a bug is notably time-consuming, taking 30 times longer than writing a line of code, leading to substantial financial costs, exemplified by the $113 billion spent annually in the U.S. on this issue. Consequently, the extensive time and resources allocated to debugging can delay product releases, impacting revenue, as research and development teams and quality assurance departments strive to manage increasingly complex software systems.

The text highlights the significant financial impact of downtime on major internet companies, focusing on examples like Amazon and Google. Amazon's 40-minute outage in August 2013 incurred a loss of $2.64 million, while Google's 5-minute downtime in the same month resulted in a $545,000 loss. The frequency of downtimes is emphasized, with a reference to a global map showing recent outages. A 2010 study by Emerson Network Power is cited, revealing an average cost of $500,000 per downtime event, with partial downtimes costing around $258,000. The text also notes that a substantial portion of downtime is spent on detection rather than resolution, and it suggests that downtime costs are expected to rise in the future.

In the rapidly evolving landscape of DevOps and continuous delivery, Log Analytics tools are becoming essential for maintaining and supporting complex cloud infrastructures. The software maintenance market, valued at approximately $140 billion annually, allocates $3 billion to log management with an anticipated growth rate of 12% annually over the next five years. As traditional log management methods become inadequate, companies are developing sophisticated visualization and search tools to filter out noise and extract valuable insights. Initially dominated by Splunk's on-premise solutions, the market saw a shift with Loggly's SaaS alternative, prompting other companies like LogRhythm, SumoLogic, and Logentries to launch their own SaaS offerings. Each company brings unique capabilities, such as Loggly's integration focus, LogRhythm's security features, SumoLogic's data aggregation, and Logentries' user-friendly interface. This burgeoning market is expected to continue expanding as software systems grow in complexity.