Currently, there are more devices connected to the Internet than there are people in the world. The Internet of Things (IoT) evolves exponentially and as a result, increases the physical representations of data accessible via Internet systems.
In view of the limitations of application memory size and their random-access nature, processors considerably lack energy-efficiency. Therefore, arises the need for dedicated hardware for IoT-generated data storage, effectively replacing software to accelerate the overall system․
Grovf offers GCache - an FPGA-based key-value store designed to efficiently store and retrieve cached data.
Fast analysis of semi-structured or unstructured data is crucial for business decision making.
The amount of information, having no standard format (images, emails, text, XML, videos, etc.) continues to grow due to the high reliance on digital content in computer systems and makes searching/analysis more complex.
According to IDC predictions, 85% of the generated data will be unstructured by 2025 and much of this will be in a textual form. As the largest data source, unstructured data becomes a large ground for analytics and deploying AI applications in the company.
To do analytical processing against unstructured textual data, companies usually confront several obstacles and need to use specific approaches to handle them, such as Regular expression algorithms.
The enormous growth in data constantly challenges computing technologies and traditional approaches we used to demonstrate. Memories barely keep pace with current tremendous growth of data and processors advancements, turning into a core bottleneck for computational systems.
Though computer technology has made incredible progress throughout the time, the dropping efficiency of modern computing points towards much more constructive changes in architecture design, known as memory-centric architecture.
Aiming to get the most out of memory-centric architecture, Grovf develops MonetX technology that expands the memory of the existing servers to several TBs per server while accelerating the near memory computing workloads.
Fast analytics of 100G network traffic on a single network card without CPU involvement is crucial for several applications, such as Lawful Interception, Deep Packet Inspection and Network Monitoring. Some applications require tens of thousands of simultaneous rules to be tracked in real-time. Others require very complex regular expressions. As a part of security solutions, Grovf offers GSearch. This probabilistic match engine is an implementation of the standard matching algorithm on the FPGA chip, achieving 100 Gbps throughput with a single IP core while supporting more than ten thousand simultaneous rules. Rules can be dynamically added, deleted, and changed on the fly.
Grovf probabilistic matching engine achieves 100 Gbps throughput regardless of the searching rule size and the number of rules (Maximum 13k rules at 100Gbps speed). In comparison, software implementations are much slower. Software speed decreases as the Rule size or a number of Rules increases.
The expansion of IoT devices, modern DevOps processes, cloud computing and encryption heavily affected the enterprise network. Increasing network complexity made it harder for security teams to imply more efficient threat detection and responsiveness. Moreover, Deep packet inspection faces challenges when processing traffic at high-speed networks.
So, security practices should evolve as well, giving priority to advanced network traffic analysis and aiming to accelerate DPI technology. Grovf offers WireHex - Deep Packet Inspection & Analysis Tool, designed for 100Gbps networks.
WireHex enables 100Gbps network real-time data analytics and the retrieval of sophisticated statistical information, reaching ~99% data visualization accuracy. Based upon the Xilinx Alveo cards, WireHex acts as a transparent network device that performs advanced network analysis, DPI and firewalling operations. It supports packet blocking based on network header parameters as well as the payload lookup using 20K user-defined rules.
The Grovf RDMA IP core and host drivers provide RDMA over Converged Ethernet (RoCE v2) system implementation and integration with standard Verbs API.
RDMA IP is delivered with reference design which includes the IP subsystem itself the 100G MAC IP subsystem, DMA subsystem, host drivers and example application on software. The system drivers are integrated with OFED standard Verbs API and are compatible with well-known RNIC cards and software. Additionally, it provides low latency FPGA implementation of RoCE v2 at 100 Gbps throughput.