Emerging from stealth mode this week, Aparna Systems announced the Orca µCloud and Orca µServer—providing an open software Cloud-in-a-Box solution. Going beyond “hyperconvergence,” Aparna says it provides “ultra-convergence” with the combination of compute, storage and networking resources in a compact, energy-efficient system. “Hyper-convergence is compute and storage with a hypervisor on it. In our case it is that plus networking. We provide the networking software and networking infrastructure,” said Sam Mathan, CEO and co-founder of Aparna.
Addressing the performance and availability needed in distributed, edge, aggregation and core environments, Aparna’s µCloud system aims to deal with what the company sees as a persistent problem in the industry caused by the way the cloud has turned server clusters into “black boxes” that make it difficult to match hardware and software to achieve peak performance. Aparna is initially offering two products- Orca µCloud system and the Orca µServers, said Mathan.
The two models of the Orca µCloud system that are currently available include the 4060 with slots for up to 60 µServers and the 4015 with slots for up to 15 µServers. Both µCloud systems are packaged in a 4U NEBS-compliant chassis with dual, hot-swappable AC/DC power supplies, and include a GPS clock to support applications that require precise timing, such as time series databases. Both systems are also equipped with dual, hot-swappable active/active switches that deliver 20 Gbps of bandwidth per µServer and an aggregate uplink capacity of 640 Gbps (2 switches x 8 ports at 40 Gbps each). This design is what gives the model 4015 its fully fault-tolerant, non-stop, non-blocking performance. The ultra-converged Cloud-in-a-Box systems offer high CPU core density (10,000 cores per rack with the 4060), energy efficiency (less than 75 watts per µServer), and high intra-cluster bandwidth.
The Orca µServers are also available in two models: the Oserv8 with 8 cores and the Oserv16 with 16 cores. Both µServers are packaged in a hot-swappable cartridge form factor that is about the size of a 3.5-inch hard disk drive. Both utilize Intel Broadwell Xeon processors, DDR RAM and dual SATA or NVMe solid state drives (SSDs). Storage I/O is non-blocking based on its support for both SATA at 12 Gbps (6 Gbps per SSD) and NVMe at 64 Gbps (32 Gbps per SSD), with latencies of 100 microseconds (µs) and 10 µs, respectively.
According to Aparna, the ultra-convergence of compute, storage and network resources in a configurable, compact server cluster enables the Cloud-in-a-Box system to deliver a savings of up to 40% in capital and operational expenditures, and to reduce the need for rack space and power by up to 80% compared to clusters built with rack or blade servers.
The Orca µCloud system is purpose-built for environments with limited availability of space, power and/or cooling.
Its open software architecture and high performance make the Cloud-in-a-Box suitable for a variety of networking, computing, or storage applications, including Fog and multi-access edge computing, databases, data analytics, the Internet of Things, artificial intelligence and machine learning.
The µCloud systema’ high density, coupled with its high performance and reliability, are expected to make them popular for edge computing and aggregation applications, as well as enterprise applications where the ability to scale compute and storage resources is often constrained by available data center space and power. In addition, said Mathan, the systems are brown field compatible, so they can interoperate with a current infrastructure.
“We are addressing multiple markets, which from a business perspective allows us to then—as we grow—manage our revenue growth, taking advantage of the disparate capital cycles that all different markets have,” said Mathan.
For more information, visit www.aparnasystems.com.