While AI training dims the lights at hyperscalers and cloud builders and costs billions of dollars a year, in the long run, there will be a whole lot more aggregate processing done on AI inference than on AI training. …
The Odious Comparisons Of GPU Inference Performance And Value was written by Timothy Prickett Morgan at The Next Platform.
Today on the Day Two Cloud podcast we're going to talk with someone who was part of a DevOps teams deploying Infrastructure as Code (IaC) and applications in the public cloud. This project ran into challenges around scaling, the environments they needed to support, how to store certain artifacts, working with pipeline, and breaking up a monolithic repo into smaller repos and the repercussions of that decision.
The post Day Two Cloud 169: Splitting Up Mono-Repositories In Infrastructure As Code appeared first on Packet Pushers.
Red Hat Ansible Automation Platform has seen wide-scale adoption in a variety of automation domains, however with edge use cases becoming more mainstream, the thought process around automation must shift from “complete a task immediately” to being able to run automation now and later, and respond to incoming automation requests from devices that are yet unmanaged.
In today’s hybrid cloud environment, automation exists in a tightly controlled and predictable space, meaning it’s easy to determine what endpoints are reachable and available for connection. In practice, this manifests as inventory syncs from our various management planes (think AWS/Azure/GCP/VMware) and then targeting the devices brought into Controller via those inventory syncs with automation. Cross connectivity shouldn’t be an issue: If we can see the device in a management plant, we can contact and automate against it. In addition, if there are exceptions to the “connectivity everywhere” model, Red Hat Ansible Automation Platform has features and functionality to help address more complex connectivity circumstances.
We can even take this automation approach one step further by pulling those management planes under the management of our automation, giving us the ability to really automate end-to-end. For example, Continue reading
Henk Smit conscientiously pointed out a major omission I made when summarizing Peter Paluch’s excellent description of how bits get parsed in network headers:
EtherType? What do you mean EtherType? There are/were 4 types of Ethernet encapsulation. Only one of them (ARPA encapsulation) has an EtherType. The other 3 encapsulations do not have an EtherType field.
What is he talking about? Time for another history lesson1.
Henk Smit conscientiously pointed out a major omission I made when summarizing Peter Paluch’s excellent description of how bits get parsed in network headers:
EtherType? What do you mean EtherType? There are/were 4 types of Ethernet encapsulation. Only one of them (ARPA encapsulation) has an EtherType. The other 3 encapsulations do not have an EtherType field.
What is he talking about? Time for another history lesson1.
Today, a change to our Tiered Cache system caused some requests to fail for users with status code 530. The impact lasted for almost six hours in total. We estimate that about 5% of all requests failed at peak. Because of the complexity of our system and a blind spot in our tests, we did not spot this when the change was released to our test environment.
The failures were caused by side effects of how we handle cacheable requests across locations. At first glance, the errors looked like they were caused by a different system that had started a release some time before. It took our teams a number of tries to identify exactly what was causing the problems. Once identified we expedited a rollback which completed in 87 minutes.
We’re sorry, and we’re taking steps to make sure this does not happen again.
One of Cloudflare’s products is our Content Delivery Network, or CDN. This is used to cache assets for websites globally. However, a data center is not guaranteed to have an asset cached. It could be new, expired, or has been purged. If that happens, and a user requests that asset, our CDN needs Continue reading
Palo Alto firewalls use the concept of a running config to hold the devices live configuration and the candidate config is copy of the running config where changes are made. A Commit operation causes the running config to be overwritten by the candidate config activating the changes.
Architecturally speaking, cloud-native applications are broken down into smaller components that are highly dynamic, distributed, and ephemeral. Because each of these components is communicating with other components inside or outside the cluster, this architecture introduces new attack vectors that are difficult to protect against using a traditional perimeter-based approach. A prudent way to secure cloud-native applications is to find a way to reduce the number of attack vectors, and this is where the principles of zero trust come into play.
With today’s multi-cloud and hybrid-cloud environments, networks are no longer restricted to a clear perimeter with clearly defined borders to defend—and cyber criminals are taking advantage of this fact by tricking users and systems into providing unauthorized access. While a lot of zero trust is focused on limiting access from users and devices, organizations are now also recognizing that in the world of distributed cloud-native applications, workloads themselves are communicating with each other and the same principles of zero trust need to be extended to cloud-native applications.
Because traditional security methods such as network firewalls rely on fixed network addresses, they are insufficient to protect dynamic, distributed, and ephemeral cloud-native workloads, which do not have fixed network addresses. They simply Continue reading