I/O virtualization is utilized by cloud platforms to provide tenants with efficient, scalable, and manageable network and storage services. The de-facto industrial standard, paravirtualization, offers rich cloud functionality by introducing split front-end and back-end drivers in the guest and host operating systems, respectively. Given this fact, paravirtualization incurs host inefficiency and performance overhead. Thus, emerging hardware virtio accelerators (i.e., SRIOV-capable devices that conform to virtio specification) with device passthrough technologies mitigate the performance issue. However, adopting these devices presents the challenge of insufficient support for live migration.This paper proposes Un-IOV, a novel I/O virtualization system that simultaneously achieves bare-metal level I/O performance and migratability. The key idea is to develop a new hybrid virtualization stack with: (1) a host-bypassed direct data path for virtio accelerators, and (2) a relayed control path guaranteeing seamless live migration support. Un-IOV achieves high scalability by consuming minimum host resources. Extensive experiment results demonstrate that Un-IOV achieves superior network and storage virtualization performance than software implementations with comparable performance of direct passthrough I/O virtualization, while imposing zero guest modification (i.e., guest transparency).

This paper highlights the use of the P4 language for the development of a Kubernetes load balancer and related network functions that address scale, security, and network performance requirements. Load balancers have multiple deployment scenarios from edge to data center clusters, including per-node application load distributions. A P4 data plane running on an Infrastructure Processing Unit (IPU) can serve as a highly performant, secure and flexible data plane for Container Network Interfaces (CNI) like Calico. Using P4, we can identify the packet headers and operator specific fields for load balancing with consistent service delivery across multi-cloud environments. Challenges like per flow monitoring, on-demand autoscaling and adding network policy ACLs (Access Control Lists) can be addressed with software and P4 data plane extensions on an IPU, eventually paving the path for modernized service mesh delivery.

Major Advantages of IPUs Separation of Infrastructure & Tenant Guest can fully control the CPU with their SW, while CSP maintains control of the infrastructure and Root of Trust Infrastructure Offload Accelerators help process these task efficiently. Minimize latency and jitter and maximize revenue from CPU Diskless Server Architecture Simplifies data center architecture while adding flexibility for the CSP

Abstract We identify three goals for the graphical user interface (GUI) of a network simulation environment: user friendliness, model re‐usability, and application extensibility. We address the user‐friendliness issue by proposing a simple yet efficient approach to setup parameters for the simulation models. We address the application extensibility issue with a two‐layer GUI architecture. The two layers are loosely coupled, and the first layer can be easily replaced without affecting the other components of the simulation environment. The concept of subnetwork is used to address the model reusability issue. Unlike most existing simulation packages, where a subnetwork is simply a method to group the nodes, the subnetwork construct in our approach is a first‐class citizen in the simulation environment (i.e., all operations for a basic node also apply to a subnetwork). The port concept is used to define the I/O relationship between a subnetwork and the outside world. Parameter dialog boxes are used to set up the parameters for a subnetwork conveniently. Finally, a simple stack mechanism is used to measure subnetwork‐wide output statistics.