Huawei will showcase its CloudEngine 12800 line of core switches, which feature a switching capacity of up to 48Tbps and bandwidth of 2Tbps per slot, Huawei says. The CloudEngine 12800 and the CloudEngine 5800/6800 top-of-rack switches support connectivity for Gigabit Ethernet, 10G Ethernet, and 40/100G Ethernet, as well as features for network virtualization and computing, storage and network convergence.
The CloudEngine 12800 series consists of three chassis-based platforms with four-, eight- and 12-I/O slots, and two slots for main processing units. On the rear side of the switch are slots for six fabric modules -- five primary, plus one for redundancy.
The line is based on a non-blocking CLOS fabric architecture, and features front-to-rear ventilation to improve heat dissipation. The high-end CloudEngine 12812 will support up to 96 100G Ethernet ports, 288 40G Ethernet ports and 1,152 10G Ethernet ports.
For network virtualization, the CloudEngine 12800 series incorporates Huawei's Cluster Switch System (CSS) feature to virtualize multiple switches into one logical switch. It also includes a feature called Virtual System (VS) to virtualize one switch into multiple independent logical devices.
Together, CSS and VS are designed to turn the network into a resource pool so that network resources can be allocated on demand.
The CloudEngine series also supports TRILL, the IETF specification for multipath Ethernet forwarding in data center and cloud fabric architectures, and will likely support other fabric architectures as well. With TRILL, network administrators can use CloudEngine switches to build large-scale Layer 2 networks with over 500 nodes, Huawei says.
For storage convergence, the CloudEngine series also supports Fibre Channel-over-Ethernet to tunnel Fibre Channel SAN traffic through an Ethernet network. The CloudEngine series also combines priority flow control with a feature called enhanced transmission selection to help ensure non-blocking transmission.
Last fall, Huawei expressed some hesitancy on entering the data center fabric switching arena, opting to wait until "competitors kill each other" over different fabric architectures before deciding to step in. But now that the battle shows no signs of letting up, Huawei's time has come, says John Roese, senior vice president and general manager of Huawei's U.S. R&D Center.
"We realized it wasn't going to settle down into one particular architecture," Roese said. "Last year, we didn't know which side to bet on. We realized it will probably be likely that there will be a number of technologies operating in concert, used for different purposes and different environments. So the extra six months gave us the time to get our technology footprint in place where we feel very comfortable whether it's TRILL, or Shortest Path Bridging, or other technologies, now we have competent capability inside of the infrastructure that can support them as feature inside the same platform."
Sign up for Computerworld eNewsletters.