News
Nebulon, Inc.® announced its next-generation Medusa2 SPU
November 2023 by Marc Jacob
Nebulon, Inc.® announced its next-generation Medusa2 SPU. It is the first and only converged infrastructure data processing unit (DPU) to unify storage, cyber and networking services on a secure, cloud-managed, server-embedded PCIe card. The Medusa2 SPU marks the first-time enterprises and service providers can unify enterprise data services, cyber and network services, and server lights-out management integration, all on a single DPU. This innovation eliminates the substantial server CPU overhead and cybersecurity risk associated with hyperconverged infrastructure (HCI) software, effectively freeing up more processor cores and equating to a 25% reduction in software licensing costs, data centre space & power consumption.
Enterprises and service providers have infrastructure that needs to remain on-premises, particularly for edge deployments, but desire the infrastructure efficiency and security developed by hyperscalers. Given the limitations of existing alternatives, more enterprises are exploring the DPU-assisted hyperscale data centre architectures for on-premises solutions based on current DPU market growth estimates, expected to reach $8.11 billion by 20281. This growth, further supported in research by Gartner2, suggests that [DPUs] “can reduce the number of servers and hypervisor licenses by 10% to 30% and may also decrease the number of application software licenses.” Gartner has also stated, “pulling security out of the server [operating system] reduces the software-based surface area for attack.” 2
The next-generation Nebulon ‘Medusa2’ SPU is a PCIe Gen 5 DPU leveraging the NVIDIA BlueField-3 Networking Platform. Similar to hyperscale approaches, the SPU establishes a secure zone in the server separate from the operating system and applications, known as the Nebulon Secure Enclave, delivering key efficiency, security, and sustainability benefits particularly when compared to HCI solutions. Storage, network, and cyber services, along with server management, are completely offloaded from server resources onto 16 powerful Arm-based processors and 48GB of memory on the Medusa2 card. This reduction in overhead on server CPUs and memory means customers can support up to 33% more application workloads per server, effectively lowering server and software license costs, as well as reducing data centre space and power requirements, by up to 25%.
The Nebulon Medusa2 SPU is operating system and application agnostic and does not require any additional drivers or software agents to be installed on the server. With the Nebulon ON cloud control plane, enterprises can benefit from the operational efficiencies offered by cloud-scale automation and fleet management of their application infrastructure—which is especially attractive for organisations with edge deployments.
Nebulon Medusa2 SPU provides unified infrastructure services on a single DPU:
• Enterprise Data Services including data reduction, data protection and NVMe over fabrics (NVMe-oF)
• Cyber Services including immutable boot/reboot from a trusted image, real-time in-line threat detection and 4-minute ransomware recovery
• NVIDIA ConnectX Ethernet NIC features and hardware acceleration
• Server Lights-out Management integration with Dell PowerEdge, HPE ProLiant, Lenovo ThinkSystem and Supermicro Ultra servers.
• Operating system support and integrations with VMware vSphere, Microsoft Windows Server and Linux bare metal/containers
