In 2005, I was invited to an evening meeting in Cupertino, California, where a group of angel investors were listening to an investor pitch from a first-time CEO of a startup company.Many in the audience were impressed by the presenter’s infectious enthusiasm, and even more impressed that each of the company’s six founders chipped in their own money to turn their collaborative idea into a product they called “a data access switch”. The meeting was such a success that at the end of the presentation, one audience member committed to writing a check to the company for a few hundred thousand dollars.
It was a critical time in the evolution of Ethernet networks. Gigabit Ethernet was just beginning to transition from a few high-cost uplink ports, to being used for the mid-2000s era of inexpensive rackmount servers. In the highest performance servers, even 10 Gigabit Ethernet was being deployed.
Five years earlier, when the Ethernet speeds used to connect servers jumped 10X, network traffic analyzers lagged behind and the few products that could sniff packets at 1G speeds were very expensive. Network IT groups would typically buy only one traffic analyzer box and strap it to a cart, borrowing a term from the carts used in hospitals, they called them “crash carts.” Network Ops would roll the cart around and plug it in to wherever the network seemed to be having the digital equivalent of a heart attack.
The startup team had an idea that was based on a couple of factors:
- First, the switch vendors had introduced features to allow the traffic on one switch port to be replicated (or mirrored) so that a traffic analyzer could see the traffic, without interrupting day-to-day operations.
- Second, thanks to off-the-shelf Gigabit Ethernet switch chips, small Layer 2 Ethernet switches were now relatively cheap.
What if they could modify the source code for the firmware of an Ethernet switch to make it look like an electronic patch panel? That way, the network troubleshooter, with a simple user interface, could connect their traffic analyzer to the switch and plug the other ports into the mirrored ports on the data switches.
One small obstacle. It wasn’t like it is today where you can buy white-box switches and just grab one of several open source network operating systems. Where were they going to get a switch and the network protocol software source code?
In a classic case of the old saying, “If we had some honey, we could have biscuits and honey, if we had some biscuits”, the startup team approached another startup where I was CEO. Our company had developed an Ethernet switch for use in a blade server, and the newer startup asked if they could use it.
Between sympathy for the startup operating on a shoestring and them having personal relationships with the investors in the blade server startup, our blade server company, with the encouragement of our investors, agreed to sell them blade server switches and provide them the source code for a fee equal to what a single cheap stackable switch would cost.
Based on that agreement, the startup raised their small angel financing and went on to create a new product category in the network market. Other companies introduced similar products, including Niagara Networks’ forerunner, Interface Masters Technologies.
Eventually, the new company stopped calling their product a data access switch and adopted a name attributed to a couple of Gartner analysts: Network Packet Broker (NPB).
A few other startups, some acquisitions, and one IPO later, several pundits predicted that the NPB market was an interesting niche, but ultimately not a very big one.
FROM NETWORK PACKET BROKERS AND BYPASS SWITCHES TO NETWORK SECURITY VISIBILITY
As time passed, data centers and service providers found a new use case for Network Packet Brokers.
To address increasingly complex security threats, data centers and service providers deployed multiple specialized security appliances to complement their firewall gateways. In linking them together one after the other, performance suffered because the slowest device in the service chain created a bottleneck that slowed down traffic to and from the data center. Even worse, because the devices were connected sequentially, upgrading the software in one of the boxes meant the whole chain had to be taken out of commission.
To avoid these issues, users figured out how to combine Network Packet Brokers with Bypass Switches and connect the security tools to them so that they could take one tool out of service without disrupting the entire system.
Today, the rise in demand for these products is attributed to security applications, and the future for most vendors looks quite bright.
WHAT COMES NEXT?
For starters, discussions about being ready to deploy 100Gb are becoming more frequent. In fact, leading edge mobile service providers are already deploying 100Gb Network Visibility products. For carriers that are committed to providing their customers the best service at competitive prices, 100Gb is key. Niagara Networks is uniquely positioned with solutions combining configuration flexibility, integrated bypass, and grooming (subdividing traffic) to lower speeds for these customers.
Beyond that, we see big opportunities to deliver increasing scale, better performance, and intuitive management.
Niagara’s Packet Brokers provide security tools with the relevant packets they need to see without becoming oversubscribed, which ultimately increases efficiency. Make your security tools more efficient and effective today with Niagara Networks Packet Brokers!