Like any pleased-with-herself three-and-a-half-year-old showing off her lace panties, start-up switch house Blade Network Technologies was flaunting its heady $230 million valuation Wednesday after closing a $10 million B round.
The largely institutional round – and let’s remember traditional VCs are ducking down any convenient rat hole these days – was led by NEC with help from Blade OEM Juniper Networks, existing Blade backer Garnett & Helfrich Capital – ex-Oracle lieutenant Terry Garnett’s venture buyout operation – and a mysterious silent investor, reportedly a big tech company that doesn’t want people to know where its money is going.
How utterly beguiling. A veiled lady in a little black dress.
The valuation was five times what the company was worth at the time of its $30 million A round in early’06.
Blade claims to leads the data center revolution for high-speed, cloud-ready networking, a proposition that, say, oh, Cisco might take exception to, but then Blade would retort that it holds a commanding 48% market share lead over Cisco in blade server networking, which, as you can guess from its name, is its specialty.
It all depends on how thin you slice the pie and if that particular slice is the only one that’s hot.
Blade is supposed to use the money to develop stuff like a 64-port 10G top-of-rack switch and software as well as for global sales and marketing for its continuing assault on Cisco.
Not that it seems to need much help. A pre-collapse Nortel Networks spin-out, Blade says that despite the economic maelstrom that took the wind out of Cisco’s sails it had a record quarter ending July 31 with revenues up over 30% when shipped its six-millionth Ethernet switch port. It claims 300 of the Fortune 500 as customers.
It claims it’s profitable and will do $100 million the next 12 months.
Of course it helps that it sells through HP and IBM – which between them own the blade market – as well NEC and Verari Systems, some of the folks Cisco alienated with its Unified Computing server thrust. Any of these boys could, of course, be Blade’s mystery investor.
Blade’s got what it calls Cloud-Ready Network Architecture solutions aimed at next-generation data centers where virtual machine awareness, ultra-high performance, extremely low latency, simplified management, lower cost and energy efficiency are key.
Its gigE switches for blade servers support multi-vendor, virtualized data center environments, promising performance coupled with lower TCO.
The trick is in the software in both its embedded and top-of-rack switches. Blade recently broke new ground with the industry’s first 1,000 port Ethernet VMready switch and the first FCoE-ready blade server switch. It’s also championed a concept called “Rackonomics,” a way of scaling data centers by duplicating racks of servers and disk drives with connectivity through its switches, your basic rack-to-go.
A few days ago Blade announced that its 1,000 virtual port VMready switch-resident software is available for HP’s BladeSystems and HP’s 1:10Gb Ethernet BL-c switch. It’s already available on IBM’s 1/10Gb Ethernet Switch Module.
The widgetry is supposed to enable virtualization across the data center infrastructure rather than just the individual servers – a technique that’s supposed to be good for energy efficiency, business continuity, productivity and resource utilization.
Being VM-aware, virtual machines can be added, subtracted and moved around, as VM are wont to do, without losing their ACLs, QoS and VLAN attributes.
According to Blade, which maintains that traditional switches are blind to VM-specific traffic, “In today’s bandwidth-intensive virtualization infrastructures, network policies must move along with virtual machines to maintain application performance, availability and security.”
VMready is supposed to manage and measure VM traffic and assign network parameters per VM, as well as maintain consistent network settings during VM migrations, allowing for workload balancing and HA failover.
Nothing changes on the server since it runs on the switch, leaving more CPU oomph for the virtual machines, and it’s all automatic.
Blade claims other advantages over both its Cisco and Arista rivals.
Unlike, say, Cisco’s Nexus 1000V software switch, its widgetry works with all the hypervisors, VMware ESX 3.x and vSphere, Microsoft, Xen, KVM and OracleVM. It says it doesn’t entail complex changes to the hypervisor, it’s standards-based, not a lock-in, it’s simpler and offers more performance and scalability. And it’s way cheaper, anywhere from $160k-$500k less than Cisco for a rack of blade servers with 128 CPU sockets running vSphere Enterprise.
Subscribe to Eclipse Developer's Journal Email News Alerts
Subscribe via RSS
