Feeds

IBM and NEC tag team on OpenFlow networking

Control freaking network configs from servers

Security for virtualized datacentres

IBM and NEC are joining forces to promote switches and controllers based on the OpenFlow protocol to their joint customers.

After a decade-long hiatus, IBM jumped back in the network-equipment business in September 2010 with its $400m acquisition of Blade Network Technologies, which made both blade and rack switches. Big Blue has aspirations in networking – as all system vendors do– and so does its systems competitor NEC, which also sells its own line of switches.

But the two companies are eager to get on the front-end of the dynamic, reconfigurable networking wave that they expect to come through the adoption of OpenFlow-based, software-defined networks, hence their partnership, announced Tuesday.

As El Reg has explained in detail, OpenFlow takes some of the management features from inside the switch and moves them to an external controller. By doing this, you create a layer of abstraction for the network (much like a hypervisor does for a physical server) that allows for networks to be defined virtually, hiding the details of the physical network and their protocols.

By doing so, you can use an external controller (analogous to a hypervisor control console) to automatically create and delete networks and give a single, virtualized view of the network. Switches have a control plane and a data plane, and OpenFlow centralizes the control plane for Layers 1 through 4 of the network onto a management appliance while leaving the data plane back on the switch.

Back in the fall, NEC launched its PF6800 ProgrammableFlow Controller Appliance, which is the first generally available OpenFlow controller. It is a rack-based x86 sever running a Linux operating system that has been hardened and equipped with the OpenFlow 1.0.0 stack. Among the many things that this OpenFlow controller from NEC can do is cope with virtual-machine live migration, automatically reconfiguring network connections when VMs flit from physical server to physical server and jump from one hypervisor instance to another.

Some switches (such as those from IBM with the BNT brand) have features to handle this live migration network reconfiguration, either with VLAN magic in Layer 2 or at the rack level in Layer 3. But the idea is that any OpenFlow-capable switch that has its flow tables being managed by an OpenFlow controller should have this capability inherently, thanks to OpenFlow, and consistently regardless of who the switch manufacturer is.

Perhaps more importantly, VM migration reconfiguration tops out at around 1,000 VMs when you are messing around in Layer 2 or 3 of the network from within a single switch, while the OpenFlow approach allows you to scale to 20,000 VMs, according to IBM.

Stewart Raphael, who used to be vice president of sales for BNT down on Wall Street, and who is now in charge of business development for the IBM's System Networking division, tells El Reg that depending on the size and complexity of the switches, the OpenFlow controller sold by NEC as either a hardware appliance or as a software appliance that runs atop SUSE Linux that can handle 15 or more switches. The OpenFlow chatter from the controller out to the switches has very small message sizes, so there isn't much overhead on the network to pull the control plane out of the boxes.

Of course, not all switches support OpenFlow yet. IBM's RackSwitch G8264, which has 48 10GbE ports and four 40GbE uplinks, and which was announced back in October 2010, was updated last November to support OpenFlow.

In Tuesday's announcement, IBM and NEC are certifying that the NEC PF6800 OpenFlow controller and the RackSwitch G8264 play nice with each other. IBM has stopped short of OEMing or reselling the controller, which has a list price of $75,000 – at least for now. The G8264 switch runs $29,999.

NEC also sells two OpenFlow-capable switches. The Univerge PF5820 looks very similar to the IBM G8264 (and may even be an OEM product, for all we know). It has 40 10GbE ports and four 40GbE uplinks, and sports the same 1.28Tb/sec of switch throughput; it costs $30,000, essentially the same price as IBM's switch. NEC also sells the Univerge PF5240, which has 48 Gigabit Ethernet ports and four 10GbE uplinks with 176 Gb/sec of switch throughput; it costs $25,000. ®

Providing a secure and efficient Helpdesk

More from The Register

next story
Docker's app containers are coming to Windows Server, says Microsoft
MS chases app deployment speeds already enjoyed by Linux devs
IBM storage revenues sink: 'We are disappointed,' says CEO
Time to put the storage biz up for sale?
'Hmm, why CAN'T I run a water pipe through that rack of media servers?'
Leaving Las Vegas for Armenia kludging and Dubai dune bashing
Facebook slurps 'paste sites' for STOLEN passwords, sprinkles on hash and salt
Zuck's ad empire DOESN'T see details in plain text. Phew!
Windows 10: Forget Cloudobile, put Security and Privacy First
But - dammit - It would be insane to say 'don't collect, because NSA'
Symantec backs out of Backup Exec: Plans to can appliance in Jan
Will still provide support to existing customers
prev story

Whitepapers

Forging a new future with identity relationship management
Learn about ForgeRock's next generation IRM platform and how it is designed to empower CEOS's and enterprises to engage with consumers.
Why and how to choose the right cloud vendor
The benefits of cloud-based storage in your processes. Eliminate onsite, disk-based backup and archiving in favor of cloud-based data protection.
Three 1TB solid state scorchers up for grabs
Big SSDs can be expensive but think big and think free because you could be the lucky winner of one of three 1TB Samsung SSD 840 EVO drives that we’re giving away worth over £300 apiece.
Reg Reader Research: SaaS based Email and Office Productivity Tools
Read this Reg reader report which provides advice and guidance for SMBs towards the use of SaaS based email and Office productivity tools.
Security for virtualized datacentres
Legacy security solutions are inefficient due to the architectural differences between physical and virtual environments.