Overview
The explosion of transporting data over cellular networks combined with the move to virtualized services creates challenges for network operators, both in meeting demand and delivering new services.
Additionally, 5G is driving a change in network architectures with O-RAN disaggregating hardware and software. 5G technology is driven by
- Up to 100GBPS data rate, which is a 10x to 100X increase from 4G and 4.5G networks
- Up to 100X number of connected devices per unit area compared to 4G LTE
- 99.999% availability, with 1-millisecond latency
- 1000X bandwidth per unit area
To offer 5G technology and services to customers and overcome challenges, telecom operators are adopting Network Function Virtualization (NFV) and Software Defined Networks (SDN) approach. Network Function Virtualization (NFV) offers network operators the ability to scale their networks with modular software running on COTS servers. HCL NFV acceleration on a SmartNIC (Network Interface Card) is a cost-effective way to improve computing capacity to accommodate new services and scale with existing server infrastructure. HCL Software has partnered with Intel to deliver a full suite of network acceleration workloads onto Intel SmartNIC and IPU cards.
Business Benefits
Optimized TCO
- Consolidate servers in your data center
- Optimized total cost of operation(TCO)
- Do more with less
Maximize Throughput
- Maximized threwput
- Mininal and predictable latencies
- Improved server utilization
Minimal and Predictable Latencies
- Up to 5X lower latency than software
- 3X CPU Core Savings
- Resources can be freed up for other business applications
Provide seamless Orchestration
- Easy to repurpose
- Programmable acceleration
- Provide seamless orchestration
Products
- 5G User Plane Function(UPF) Acceleration
- vRouter Acceleration
- TF vRouter Acceleration
- OvS Acceleration
- SRv6 Acceleration
5G User Plane Function(UPF) Acceleration
5G UPF is a critical function in 5G operators’ network which specifically requires high performances in terms of throughput, latency, and flow insertion rate for subscriber mobility support. HCL provides a DPDK-based solution that leverages Intel FPGA SmartNIC N6000-PL Platform. HCL Solution can be easily integrated with any UPF Software Stack and provides outstanding performance:
- Upto 16 Million Flow entries at HW
- Upto 200G Throughput
- Ultra Low Latency – less than 4 microseconds
- Flow insertion Rate - Upto 300K flows per second
- Zero cores used for fast path
- Kubernetes Orchestration
vRouter Acceleration
vRouter Acceleration is a software-defined networking (SDN) support solution that help automate the creation and management of virtual networks. It lets the user connect, isolate, and secure workloads in both private and public clouds. This solution can be Integrated with OpenStack, OpenShift, and Kubernetes and provides support for hybrid SDN orchestration and centralized control of virtualized switching, routing, security, load balancing.HCL Solution running on Intel FPGA SmartNIC N6000-PL Platform provides:
- Flow based and Packet based support for 50G/100G solutions with rich feature set
- Zero cores used for fast path
- Kubernetes Orchestration
Tungsten Fabric (TF) vRouter Acceleration
Tungsten Fabric (TF) is an SDN hybrid and multi-cloud solution which allows secure and unified policy enforcement across the cloud environment. TF vRouter Acceleration improves the vRouter forwarding rate using Intel’s PAC N3000 card, freeing CPU resources for VNF/CNF use.
- 100% core savings for fast path (zero cores used for fast path)
- Up to x6 times increase in throughput for smaller packets & Up to x5 times lower latency than software
- Up to 100% line rate for higher packet size
A teleco -grade OvS acceleration solution for both VM and Containers-based infrastructures
Open vSwitch (OvS) is widely used in SDN and NFV communities. Despite the flexibility of virtual switches, they bring poor I/O performance and high CPU overhead whereas OvS Acceleration resolves performance issues. HCL OvS acceleration solution improve performance and save CPU Cores. It supports both OVS-Kernel & OVS-DPDK implementation and has robust features set.
- Achieve 10X performance improvement with HCL’s OvS acceleration solution
- Supports both OVS-Kernel/DPDK implementation and has a robust feature set
- Migrate workload in real-time (scaling) and achieve live mirroring
Segment routing over IPv6 addresses the requirements of NFV & SDN architecture
SRv6 simplifies NFV implementation by allowing SDN, service chaining, and tunneling at the same time
- 3X performance improvement with SRv6 accelerated solution
- Save CPU cores and CPU cycles to get the same throughput as SRv6 software-based approach
- Available CPU capacity enables optimized processing of core network functions
