The Demise of Agile
How Agile doesn’t scale or handle complexity.
Credit to SOUMEN SARKARSOUMEN SARKAR on LinkedIn.
Agile in the Cloud
Agile was once hailed as a revolutionary approach to the development of software in the cloud, but is now facing rejection from enterprises. On paper, agile should be furthering innovation which makes it hard understand the frequent layoffs in the tech industry.
Why is this a problem? Layoffs create a toxic culture, leading to low morale and high stress. Agile requires high engagement to solve complex problems, but layoffs undermine this.
In 2023, Capital One laid off 1,100 workers from their “Agile” delivery division. This move suggests that Agile’s value-to-cost ratio is being questioned.
Cisco Systems also announced two rounds of layoffs in 2024, i.e., 4,000 jobs cut in February and 7% of employees (approximately 5,900 jobs) cut in August. Cisco’s acquisition strategy and focus on “synergies” probably lead to mass layoffs.
Mass layoffs also make cloud-base technologies vulnerable to revenge attacks from ex-employees. A notable example is the case of Sudhish Kasaba Ramesh, a former Cisco employee who, after being forced to resign, damaged the company’s network, deleting 16,000 WebEx Teams accounts and costing Cisco $1,400,000. On December 9, 2020, Ramesh was sentenced to 24 months in prison, followed by one year of supervised release and a fine of $15,000.
Impact Engineering — Book
Dr. Junade Ali, a renowned expert, argues that Agile’s emphasis on speed and feature delivery often compromises software reliability, leading to catastrophic failures.
Ali attributes software disasters to flaws in the initial requirements stage, which cascade throughout the development process. He emphasizes that the problem lies not in testing, but in the initial design and fear of speaking up among engineers.
His book can be found here.
The CrowdStrike Incident
Agile development methodologies might not be suitable for complex systems like the Linux kernel. With around 30 million lines of code, making changes requires familiarity with the existing code and an acceptance from the community.
This would pose a significant challenge if using Agile because to the Linux kernel is a difficult process. Even if a change is accepted, it can take years to arrive in production environments due to the slow adoption of new kernel releases by Linux distributions like Red Hat Enterprise Linux. This is in direct opposition to Agile.
Agile development is not effective for non-trivial systems that matter.
Following the CrowdStrike’s 2024 incident, Microsoft posted a blog post called Windows resiliency: Best practices and the path forward. It mentions security innovations like VBS enclave feature and Microsoft Azure Attestation (MAA) service. You can see below that these are kernel or operating system level features.
VBS utilizes the Hyper-V hypervisor to create an environment that is higher privileged than the rest of the system kernel. Like VM isolation, the hypervisor sets memory protections in the second level address tables and IOMMU tables to isolate this environment from the rest of the system kernel.
Azure Attestation enables cutting-edge security paradigms such as Azure Confidential computing and Intelligent Edge protection. Customers have been requesting the ability to independently verify the location of a machine, the posture of a virtual machine (VM) on that machine, and the environment within which enclaves are running on that VM. Azure Attestation empowers these and many additional customer requests.
CrowdStrike recently issued a Root Cause Analysis (RCA) of a faulty software update that led to a global IT outage, resulting in a $50 billion market cap loss. The culprit? A simple “Off-By-One Error” (OBOE), a mistake often taught to first-year programming students to avoid. This led to an out-of-bounds memory read, resulting in a system crash and the infamous Blue Screen of Death.
Microsoft’s new low-level approaches will, by necessity, prioritize security and trustworthiness over Agile development.
Agile Doesn’t Scale
Agile’s emphasis on thin slicing and marginal benefits can result in a scale of change that may not align with the original goals.
Alfred Kahn’s concept of “tyranny of small decisions” describes how incremental, near-term decisions can lead to unanticipated consequences.
Consider when the Ithaca, NY passenger railway was the only reliable transportation option, but it was withdrawn due to insufficient revenue. Despite providing year-round service regardless of conditions, the railway struggled financially. Local airlines and buses cherry-picked profitable routes, leaving trains to cover less desirable ones. Collective individual decisions by travelers, prioritizing convenience over loyalty which ultimately led to the railway being shutdown.
This discrepancy arose from differing time perceptions: travelers made short-term decisions, while the railway made long-term decisions. The cumulative effect of small decisions made by travelers ultimately led to the railway’s withdrawal. Each individual choice seemed negligible, but collectively, they voted against the railway’s continued existence.
Conclusion
- The accumulation of small decisions can lead to a “discontinuity” like sudden layoffs or a mismatch between individual goals and company objectives.
- Enterprises, as complex systems, are vulnerable to market fluctuations and interest rates, making daily standups and sprint goals insufficient for achieving collective success.
- Engineering often requires both simulation and prototyping before manufacturing
- Agile deviated from its original purpose and remove the voice of engineers
Agile’s focus on small steps and marginal benefits can be detrimental to enterprises. It’s time to reevaluate the value of Agile practices and consider a more holistic approach to software development.
