Summary:

Compass-EOS is an Israeli company that is trying to rethink routing. It has invented a new optical chip to do that. And all that means it has needed money – about $160 million in total, of which $42 million came just recently.

Compass-EOS icPhotonics - In Hand

In an era where venture capital industry is busy gurgling over ephemeral social media platforms and me-too e-commerce apps, it is nice to hear about a company that has the gumption to try and to do something big, hairy and audacious, like build a brand-new class of routers to take on incumbents like Cisco Systems and Juniper Networks

Compass-EOS (aka Compass Electro Optical Systems), which isn’t going to win any style points from brand experts, on November 18th, 2013 announced that it raised $42 million in new financing from existing investors Comcast Ventures, Cisco, Pitango Venture Capital, Benchmark Capital, Northbridge Venture Partners and Marker and RUSNANO. This new round of funding brings the total funding invested in the Netanya, Israel-based company up to $160 million — not something to sneeze at. (We call such companies phat startups!)

The company has needed every penny of it. Compass-EOS started with the idea that service provider networks were getting too crowded with specialized routers — core routers, edge routers, access routers and aggregate routers. The situation, was much like data centers were 20 years ago — over run with specialized servers, taking up space and sucking down power and in the process creating an operationally inefficient and costly infrastructure. The rise of virtualization and newer interconnect technologies have led to a different model of web-scale data centers that takes advantage of lower cost and denser servers.

Compass-EOS founders Michael Laor  and Michael Mesh thought that service provider routing infrastructure needed to adapt similar approach. It was an insightful approach — the networks today are trying to scale up to keep up with our unending thirst for bandwidth. A decade ago, 100 gigabit per second networks sounded pretty cool. Today, there is talk of terabit and beyond networks. The routers of today, like Marlon Brando – getting fatter by the day. In order to make the routers scale to super speeds, you need more silicon and in order to power those chips and all the paraphernalia around them, you need a lot more energy.

Slide15

Laor (formerly of Cisco) and Mesh (formerly of Packetlight) realized that this growth in size isn’t going to be feasible forever and started looking for alternative approaches. They started the company in 2006 and very quickly realize that in order for them to make their vision come true, they would need to tap into the fast changing world of silicon photonics. The company calls their technology, icPhotonics.

The company built a icPhotonic chip — essentially a chip than can process data at 1.34 Tbps. The chip at its center has a matrix of lasers that are used to transmit data optically at 1.34 Tbps and the chip also has photoreceptors that allows the chip to receive data at 1.34 Tbps. As a result, chips, which sit on a line card, can be interconnected with passive optical connections.

Having built the technology, Compass-EOS launched its r10004 router in March 2013. My colleague Stacey Higginbotham wrote about the commercial shipment of their icPhotonics chip-based routers, which are currently being used by NTT Communications and major research and education organizations including CERNET have successfully deployed the company’s routers.

According to Compass-EOS vice president of marketing, Asaf Somekh, the company has also sold its routers to another large global carrier, and another company in China. More customers are likely to announced soon, he told during the course of a conversation with us. They like that the company is able to replace massively huge routers such as CRS-1 with much more svelte devices from Compass-EOS, which in turn frees up space to add gear for other money-making services such as content distribution, Somekh explained.

Things have not gone as smoothly as they could for the company, but it seems with the rise of terabit networks, it seems to be in sight of finding what the web folks call, a “product market fit.” And when it does, it might be time for investors to exhale — $160 million is a lot of money to bet on a company. But so is investing $336 million in a furniture e-tailer. Well, at least one of them is a tech company!

Slide16

 

Comments have been disabled for this post