Learn how Palo Alto Networks is Transforming Platform Engineering with AI Agents. Register here
Register now
Optimize compute, storage and data
Choose copilot or autopilot execution
Continuously improve with reinforcement learning
%20System%20(640%20%C3%97%20360px).png)
Is Tesla known for building the safest car on earth? They made an interesting discovery regarding task division. By examining all operations, they identified two distinct segments: One that can be effectively performed by machines and another that is better suited for humans. When the first segment is completely handled by machines without any human intervention, it is referred to as an autonomous system. This means that the system operates autonomously.
Now, how does an automated system differ from an autonomous system?. If Tesla were to provide alerts for every left and right turn and you, as the driver, make those turns based on the alerts, that would be an automated system. In this case, the responsibility is shifted to the driver rather than relying solely on the machines. The control is taken away from the machines. An autonomous system, on the other hand, occurs when the machines themselves handle the left and right turns without requiring human input. You can watch the original video here.
The most crucial aspect of an autonomous system lies in its ability to assume 100% accountability. There is no room for a 99% accountability—it must be an unwavering 100%. This distinction carries significant importance.
To achieve autonomy, various factors come into play. Learning, acquiring knowledge, adapting to new situations, evaluating regressions, taking decisive actions, and measuring efficacy are all essential components in assuming accountability. Building an autonomous system hinges on these fundamental principles. Tasks such as system restarts, reboots, replacements, scaling up or down, whether horizontally or vertically, should be entrusted solely to machines. This allows your teams to focus their efforts on higher-order responsibilities, such as selecting the appropriate language or architecture. These challenging and intellectually stimulating tasks require human involvement and contribute to a sense of fulfillment. They are the tasks that truly showcase intellectual quotient (IQ). Consequently, your teams should not be disturbed during the night to handle automated operations.
Lately, we have been actively engaging in a process of shifting operations earlier or later in the development cycle, commonly known as shift left and shift right. Let's take testing as an example. Traditionally, testing was typically carried out in later phases or even in production. However, a new approach called testing in production has emerged, allowing for the delay of testing until the production environment. Additionally, configurations, such as Infrastructure as Code (IACs), can be handed over to developers for their configuration expertise, rather than solely relying on other teams. It's important to emphasize this key aspect we may be overlooking.
Before embarking on any shift left or shift right practices, it is imperative to initially shift up. This entails identifying all the tasks and operations that can be efficiently executed by an autonomous system, and entrusting them to that system. Then, the remaining tasks can be divided for shift left and shift right approaches. By following this approach, we create a stable foundation where the allocated operations remain fixed, eliminating unnecessary oscillation and ensuring a smoother development process.
We emerged from our stealth phase half a year ago and were recently acknowledged as a Gartner Cool Vendor two months ago. Today, we have two significant announcements. Firstly, starting from last Wednesday, July 27th, we are now an ISVA, meaning we have formed a close partnership with Amazon AWS and will co-sell with them. Secondly, we are introducing a new feature called autonomous concurrency, which addresses the primary issue of cold starts in serverless computing. Cold starts occur when a Lambda function receives a request for the first time and takes time to warm up before serving it. The challenge lies in not knowing when important client requests or seasonal demands will occur, requiring timely service delivery.
For the past seven years, a method called simple warmups has been deployed, executing warm-ups every 15 minutes to ensure the Lambda functions are up and running. However, this approach has been limited by allowing only one Lambda to run at a time. Consequently, if there are two customers, the second customer still experiences a cold start.
Today, Sedai is unveiling a new feature called autonomous concurrency. This feature is designed to eliminate cold starts by more than 99%. The goal is for you to no longer encounter cold starts when deploying Lambdas. We address seasonality at a local level and utilize edge intelligence to determine when and what needs to be warmed up. For further insights, watch the video here.
April 7, 2024
August 9, 2024
%20System%20(1920%20%C3%97%201080px).png)
Is Tesla known for building the safest car on earth? They made an interesting discovery regarding task division. By examining all operations, they identified two distinct segments: One that can be effectively performed by machines and another that is better suited for humans. When the first segment is completely handled by machines without any human intervention, it is referred to as an autonomous system. This means that the system operates autonomously.
Now, how does an automated system differ from an autonomous system?. If Tesla were to provide alerts for every left and right turn and you, as the driver, make those turns based on the alerts, that would be an automated system. In this case, the responsibility is shifted to the driver rather than relying solely on the machines. The control is taken away from the machines. An autonomous system, on the other hand, occurs when the machines themselves handle the left and right turns without requiring human input. You can watch the original video here.
The most crucial aspect of an autonomous system lies in its ability to assume 100% accountability. There is no room for a 99% accountability—it must be an unwavering 100%. This distinction carries significant importance.
To achieve autonomy, various factors come into play. Learning, acquiring knowledge, adapting to new situations, evaluating regressions, taking decisive actions, and measuring efficacy are all essential components in assuming accountability. Building an autonomous system hinges on these fundamental principles. Tasks such as system restarts, reboots, replacements, scaling up or down, whether horizontally or vertically, should be entrusted solely to machines. This allows your teams to focus their efforts on higher-order responsibilities, such as selecting the appropriate language or architecture. These challenging and intellectually stimulating tasks require human involvement and contribute to a sense of fulfillment. They are the tasks that truly showcase intellectual quotient (IQ). Consequently, your teams should not be disturbed during the night to handle automated operations.
Lately, we have been actively engaging in a process of shifting operations earlier or later in the development cycle, commonly known as shift left and shift right. Let's take testing as an example. Traditionally, testing was typically carried out in later phases or even in production. However, a new approach called testing in production has emerged, allowing for the delay of testing until the production environment. Additionally, configurations, such as Infrastructure as Code (IACs), can be handed over to developers for their configuration expertise, rather than solely relying on other teams. It's important to emphasize this key aspect we may be overlooking.
Before embarking on any shift left or shift right practices, it is imperative to initially shift up. This entails identifying all the tasks and operations that can be efficiently executed by an autonomous system, and entrusting them to that system. Then, the remaining tasks can be divided for shift left and shift right approaches. By following this approach, we create a stable foundation where the allocated operations remain fixed, eliminating unnecessary oscillation and ensuring a smoother development process.
We emerged from our stealth phase half a year ago and were recently acknowledged as a Gartner Cool Vendor two months ago. Today, we have two significant announcements. Firstly, starting from last Wednesday, July 27th, we are now an ISVA, meaning we have formed a close partnership with Amazon AWS and will co-sell with them. Secondly, we are introducing a new feature called autonomous concurrency, which addresses the primary issue of cold starts in serverless computing. Cold starts occur when a Lambda function receives a request for the first time and takes time to warm up before serving it. The challenge lies in not knowing when important client requests or seasonal demands will occur, requiring timely service delivery.
For the past seven years, a method called simple warmups has been deployed, executing warm-ups every 15 minutes to ensure the Lambda functions are up and running. However, this approach has been limited by allowing only one Lambda to run at a time. Consequently, if there are two customers, the second customer still experiences a cold start.
Today, Sedai is unveiling a new feature called autonomous concurrency. This feature is designed to eliminate cold starts by more than 99%. The goal is for you to no longer encounter cold starts when deploying Lambdas. We address seasonality at a local level and utilize edge intelligence to determine when and what needs to be warmed up. For further insights, watch the video here.
.svg)
.svg)
