You can share it to WeChat via the Qr code.
Enter the applet sharing event using WeChat scan.
From November 16 to November 21, Intel will hold packet processing virtual summit, which will be live online and divided into NIC, platform, security, workload and edge sessions. Technical experts from Intel and its partners will bring you a technology feast.
There are technical Q & A sessions, round table discussion sessions and exquisite raffle gifts. Interested users are welcome to sign up online.
Important reminder
After filling in the online form, users who sign up for the event should join the wechat group (scan the QR code below) to facilitate the organizers to synchronize live links and attendance reminders within the group in time.
Consultation: Ms. Mo,17714204576.
E810 DPDK PMD Introduction
As Intel Ethernet 800 Serial Network Controller comes into the market, the DPDK driver plays an important role that helps our customers to maximal leverage the hardware capability to benefit their workload. In this presentation, we will give an overview of latest DPDK "ice" driver. Besides the basic capability, we will go through a couple of advanced features, such as Device Configure Function for NVFi, Advanced AVF for VNF. We will also introduce abundant flow offloading features powered by Intel DDP and explain how this helps to accelerate the vSwitch by offloading mega flows.
Device Virtualization in DPDK -iAVF Backend
With NFV moving towards cloud-ready stage, it requires ideal network IO virtualization solutions which are able to meet the requirements of virtualization and cloudification simultaneously. iAVF (Intel® Ethernet Adaptative Virtual Function) backend provides software equipment for iAVF NIC(network interface card). iAVF hardware NIC provides high performance which meets the requirement of cloudification, and iAVF backend enables iAVF platform independent which meets the requirement of virtualization.
CPU dynamic load balancing module -- new eventdev device from Intel
Intel has added a new CPU dynamic load balancer (DLB) to the recently released SOC processor, which can effectively improve the utilization rate of each core when using multi-core CPU for packet forwarding. At the same time, it provides a new solution for the network elephant flow (single stream and large traffic) that the network card cannot handle flexibly. This topic will start with the working principle of DLB and introduce the corresponding software development methods.
Intel Big Spring Canyon (BSC) SmartNIC for Cloud Data Center
SmartNIC gets more and more attractions for cloud data centers, freeing up processor resources and simplifying management of data center infrastructure, especially for bare metal cloud. In this presentation, we will introduce a new FPGA based SmartNIC product Big Spring Canyon (BSC), which accelerates networking (e.g. vSwitch), storage and security workloads. Then we will introduce how DPDK rte_flow/rte_mtr those common APIs offer customers the ability to customize their solutions while still delivering greater levels of performance and programmability.
From low level crypto instructions to high performance IPsec/TLS : a full coverage of secure transportation acceleration
We will present Intel crypto underlying acceleration technologies for PKE, symmetric crypto and Hash etc and then introduce how to accelerate secure transportation protocols such as IPsec, TLS, QUIC based on it.
Towards Cloud-friendly IPsec Elephant Flow: VPP Asynchronous Crypto with QAT or Multi-core SW scheduler engines
Elephant flow has been an obstacle to IPsec application due to the huge cycle cost spent on crypto as well as SEQ update/check synchronization problem when trying to load balance the processing of the same flow to multiple CPU cores.
In this talk we will introduce our recent work towards conquering the IPsec elephant flow problem in VPP. The talk includes:
- The VPP asynchronous crypto infrastructure that allowing offload crypto workload.
- The new VPP Cryptodev engine that allows the user to take advantage of fast QAT symmetric key encryption/decryption.
- The new VPP multi-core sw-scheduler crypto engine for the user to assign CPU cores to do dedicated crypto processing.
Introduction to HDSLB: High Density Scalable Load Balancer
In this talk, we will present HDSLB: High Density Scalable Load Balancer.
Based on DPVS project, we refactored the framework and optimized key features to boost performance by leveraging advanced NIC features and CPU capabilities.
We also added some new features to meet customer's requirements, including Session Sync.
Finally, we will discuss the performance gain and next step for HDSLB.
Advanced features of Intel NIC and 5G UPF Acceleration based on DDP
More and more applications and workload are emerging, which need more features and performance of Ethernet controller. Introducing Ethernet technology to optimize typical of industry application. Solution for Cloud, virtualization, low latency and storage related technology will be highlighted.
Especially based on advanced DDP feature of FVL/CVL, show how to leverage enhance packet processing engine to implement packets meta data parser/extraction, and flow mark and classification to satisfy demanding 5GC UPF requirements and user scenario like eMBB and URLLC.
Reference design and testbed practice of Intel openness edge computing platform
It introduces the latest version capability of openness edge computing platform reference design, and the specific practical cases of 5g network element and edge application deployment based on 5g + MEC test bed.
Panel Discussion
E810 DPDK PMD Introduction
Device Virtualization in DPDK -iAVF Backend
CPU dynamic load balancing module -- new eventdev device from Intel
Intel Big Spring Canyon (BSC) SmartNIC for Cloud Data Center
From low level crypto instructions to high performance IPsec/TLS : a full coverage of secure transportation acceleration
Towards Cloud-friendly IPsec Elephant Flow: VPP Asynchronous Crypto with QAT or Multi-core SW scheduler engines
Introduction to HDSLB: High Density Scalable Load Balancer
Advanced features of Intel NIC and 5G UPF Acceleration based on DDP
Reference design and testbed practice of Intel openness edge computing platform
Panel Discussion