Understanding Tri-O-Core

Let’s see what is the our Tri-O-Core made of and how powerfull it is.

Why performance Matters ?

Cloud computing applications and mobile apps demand high capacity multi-core or multi-processor architecture to perform seamlessly.


Applications are expected to securely deliver performance and scale to operate on any kind of environment and respond quickly at wire speed.


Nubewell Tri-O-Core architecture, harnesses the advantage of multi-core design and delivers high throughput, visibility in to encrypted traffic, Scalability, and agility at lower bandwidth with low latency.

What is Our Tri-O-Core

Tri-O-Core is Nubewell’s next generation hardware architecture allows up to 80Gbps of packet processing and 40Gbps of SSL throughput.


Tri-O-Core architecture uses a unique technology called copy-free packet processing to increase HTTP, TCP and SSL/TLS performance significantly and processing at wire speed.


Tri-O-Core allows Nubewell to deliver reliable, high throughput and full-featured network solution. Tri-O-Core architecture is optimized, efficient, scalable, and flexible, which can take on any load without compromising on quality, price, or performance.

Datapath

Tri-0-Core’s DataPath provides packet processing, prioritizing and application classification.



Copy-free Packet Processing

Once the data packet received from wire, there is no copy or duplicate of the packet at any layer for packet processing until the packet is transmitted out.


NUNK Operating system

NUNK is “No User-space No Kernel-space” operating system. NUNK OS runs in standalone mode and provides low overhead and a high potential for scale on multi-core environment for packet processing.


Tri-O-Core Data Path takes care of network resources by prioritizing data packets in the hardware. Hardware packet processing has many advantages over software packet processing. It offers high performance and high QoS accuracy in hardware packet processing whereas, software packet processing performs slower due to additional processing of context switch.




Coherent multi-core and multi-pipeline hardware scheduler

Main memory is shared across all the cores which is primary communication for bulk transfer of data traffic. A set of cores can be configured as core group to perform pipelining for different data plane operations which will make sure all the packet processing happens in parallel. Different sets of cores may execute different portions of data plane services.


Hardware-accelerated packet parsing and packet allocation

All the packet parsing functions are implemented in the inline hardware to offload from the cores for all the data movement. When a packet hits the core for software processing, the packet parsing, packet memory allocation and other calculations are performed at the hardware. All the packet buffer allocation and deallocation are automatically done by the hardware.



DPI protocol and application classification

The Deep Packet Inspection protocol identifies packet sizes, timings, ratios, byte patterns, etc., and classifies applications based on heuristics.
This DPI Engine can identify Layer 4 to Layer 7 classification for more than 2,800 applications and protocols. It can support more than 6,000 metadata parameters.

Crypto Accel

Crypto Accel block is the high performance encryption mechanism employed by Tri-O-Core architecture.



High performance security processing

The crypto acceleration block processes packets up to 40G of SSL and IPsec. A self-contained operating system makes sure that performance is not affected during Encryption and Decryption.


Wide variety of algorithms supported

Nubewell crypto path Supports a wide variety of protocols including IPsec, SSL, TLS 1.2 including several variants of AES, 3DES, SHA-2, SHA-3, RSA 2048, RSA 4096, RSA 8192, ECC p256/p384/p521.,


Nubewell's hardware contains compression/decompression (ZIP) engines that support DEFLATE/ZIP and LZS algorithms.


FIPS 140-3 compliant with True RNG

Compliant with US government's latest computer security guideline for cryptographic modules. Includes a True Random Number Generator which produces highly unpredictable random numbers for encryption.



Bulk encryption/decryption with symmetric/asymmetric Engines

Our encryption algorithm makes sure that sensitive and non-sensitive data is identified using SSL visibility and data packets are selectively chosen for symmetric or asymmetric cryptography. High performance chips make sure that data compression/ decompression is done without data loss.


Non-blocking crypto packet processing

Non-blocking algorithm makes sure that packet encryption is not affected when there is interruption or packet loss.

Control Path

Control Path core takes care of visibility, Analytics and industry leading CLI Support with multi tenent architecture.



Industry Leading Network CLI support

We use CLI which is widely adopted by the Network switch CLI makes sure that interacting with the appliance for all the network solutions and analytics is easier. Our CLI is friendly for programming and automation tasks.


High Availability Architecture

High capacity system-on-a-chip processors used in Tri-O-Core makes sure that high availability is maintained. High availability architecture is durable and fault-tolerant. High availability provides continuous uptime for load balancing, backup, and continuous packet processing.


Nubewell's hardware contains compression/decompression (ZIP) engines that support DEFLATE/ZIP and LZS algorithms.


FIPS 140-3 compliant with True RNG

High capacity system-on-a-chip processors used in Tri-O-Core makes sure that high availability is maintained. High availability architecture is durable and fault-tolerant. High availability provides continuous uptime for load balancing, backup, and continuous packet processing.


Multi-tenant/NFV

Multi-tenant architecture makes sure that resources are shared across applications and each tenant's data is kept privately for each user or application. Our Network Function Virtualization (NFV) technique helps you run different virtual machines on the existing infrastructure instead of having dedicated hardware.


Big Data Analytics Engine

Big data analytics helps find out hidden patterns, process huge traffic data, and get useful business intelligence.


Open Flow/NETCONF

Our Control Path supports OpenFlow and NETCONF protocols which are the norm in Software-Defined Network infrastructure.

OpenFlow protocol provides an open architecture in which control functions are separated from the network device and placed in accessible control servers. With OpenFlow, packet moving is centralized allowing for greater programmability. Routing decisions can be made periodically or at a user-specified time.

NETCONF protocol helps you install, manage, and delete the configuration of network devices. NETCONF protocol uses an XML to save network configuration details and messages. XML is a highly structured markup language which helps handling of configuration data and messages easier.