#north_bridge

After years of social media startups getting all the love, enterprise software is back. Last month, GigaOM said we’re living “in the golden age of enterprise startups.” Now comes Gartner with a new report

saying enterprise software buys are helping accelerate global IT spending 2X over 2012, to nearly US$4 trillion. With spending back up, more and more entrepreneurs are pitching VCs – including this one – to fund enterprise software startups. And yet, as someone who spent 15 years as a senior software sales executive, co-founder, and CEO before going into venture capital, I’m amazed how many startup entrepreneurs try to get into the enterprise – and into VCs’ portfolios – without following the Three Golden Rules of Enterprise Sales:

1. BUILD WHAT YOU SELL, DON’T SELL WHAT YOU BUILD. Every successful software entrepreneur can tell stories of the dark early days when they stupidly thought customers really needed X, but soon found out they actually wanted Y. Founders get so passionate about their brilliant ideas they often forget to go ask the enterprise what it really needs. Are the pain points really there? Is the workflow friction as bad as you think? The No. 1 priority here is rooting your value proposition in a rock-solid ROI proposition atop a diamond-hard business case. Your solution must deliver a measurable delta in revenue generation or cost reduction, profitability, and ultimately you must identify shareholder value.

2. SEEKING  VITO: UNDERSTAND THE ENTERPRISE ORGANISM.  Always talk to the most senior person who’ll see you – what Anthony Parinello called a “VITO” – a Very Important Top Officer – in his 1999 book “Selling to VITO.”  Find out how they buy and why they buy. Find out what you can sell (see Rule 1) by developing enterprise situation fluency: who has budget, who says yes, who says no (and how you can block the person who says no), who’s going to be your biggest internal sponsor – hopefully, a VITO. Talk to other companies who’ve sold to the organization you’re targeting to find out what works. Don’t make a single sales call until you’ve earned a black belt in your product, you’re steeped in its ROI or business case and you’re immersed in each customer’s unique business culture. Develop what we call “situational fluency” – identify your key sponsors in the line of business and in the IT organization.  Who owns the budget? Who has access to VITOs or has the power to influence and shepherd your agenda?

3. FAIL FAST – IN THE FAT PART OF THE MARKET. Just ship, baby! Don’t agonize over making it perfect, because it won’t be. There’s almost no chance V1.0 is going to hit the precise sweet spot, so sell it, build it, install it, then tweak it. Done is better than perfect. Customer enhancements and wish-lists generated after you ship V1.0 are the best ways of figuring out how to move versions 2.0, 3.0, and so on ever-closer to the bull’s-eye. And for God’s sake, go after vibrant markets where appetite for change is voracious. Financial services and retail, for example, are always hungry for the next bright, shiny, profitable object. You’ll have a better chance of persuading the VITOs there. Electric utilities? Not so much.

My last post described some of the market trends leading to a perfect storm in networking –

o   4x growth in the number of internet users (900M to 3.7B; 2008 – 2016)

o   5x growth in the number of devices connected to the Internet (9B to 50B; 2012 -2020)

o   6x more traffic inside data centers to assemble each page served

o   10x growth in the number of web queries (3B to 320B; 2008 – 2016)

o   10x growth in capacity requirements (1 Tbps to 10 Tbps; 2015 – 2020)

o   25x increased activity per user (2008 – 2016)

o   26x growth in mobile data (.25M to 6.5M of terabytes/month; 2010 – 2015)

o   An explosion in data -- from the dawn of civilization until 2003, humankind generated five exabytes of data. Now we produce five exabytes every two days … and the pace is accelerating.

– together resulting in a need for massive amounts, up to 500x, of additional networking capacity!

Sorry Cisco but the answer to scaling these networks to new heights is not to simply buy metal but to approach the problem in completely new ways.  To steal Steve Jobs’ famous line, networking vendors need to “Think Different”.

The solution to this looming data crisis lies in an area most had left for dead a decade ago – one word, optics.

OK, more accurately fiber optics.  But why?  A single strand of fiber can carry large amounts of data over long distances.  In the late 90s, VCs and entrepreneurs made many fortunes in supplying the optical equipment to build the backbone of the Internet.  Fiber optics made sense because the other viable alternative, copper wires, couldn’t carry enough data far enough.  The cost of this optical equipment was very expensive but they were building mankind’s greatest invention since the printing press – it was a long-term investment.

The combination of the collapse of the Internet bubble 1.0 and a free-fall in bandwidth prices resulted in a bloodbath in the ’01-’03 timeframe at which point optics became a 4-letter word. Now, a decade later, optical networking is back with a crucial new role to play. 

Optics are both more power efficient than copper and can carry far more capacity.  Using wave division multiplexing (WDM) – a single stand of optical fiber can carry dozens of different colors of light each of which can support 10, 25 even 100 gigabits of data.  Earlier this week, the IEEE standards organization started a working group on 400Gbps Ethernet to address the booming demand for bandwidth – which they estimate at a stunning 58% CAGR.  In total, many terabits of capacity per fiber will support the “east-west” traffic inside a data center.

Unlike last time around, when optical networking centered on long distance transmissions and networks for cities, today’s opportunity is happening inside the datacenter. As companies move more online they’re struggling to build datacenters big enough to keep pace with demand. Further, they’re hitting the limits of copper-based connectivity in terms of capacity, power consumption and cost.

But the transformative impact of optics extends beyond the datacenter to:

Server racks – fast, low power optical interconnects will allow equipment vendors to split servers into separate components – so processors, memory and flash can be upgraded and scale independently of one another

Chip-to-chip communications – as bandwidth speeds increase to 100s Gbps copper trace on a printed circuit board is reaching its redline so interconnects using high-speed optical connections are already being used in routing systems and will make their way into processing and memory trays.

On-chip optical functionality, referred to as silicon photonics, will enable optical functions (modulate, add, drop, detect) to be built alongside the transistors on the top of the chip itself. A single semiconductor manufacturing process can make both transistors and optical circuits at the same time.  Doing so will dramatically reduce the cost of the optical functions and increase their widespread use. 

Whether it’s a perfect storm or a tsunami, the massive growth in Internet activity demands a rethink in how we connect elements of the Internet, whether its coast-to-coast or chip-to-chip.  We think there’s one word you need to know to find the answer – optics.

My name’s Jamie and I’m addicted to data.  Happily I might add.  I’ve been using since my parents gave me an Apple II as a child and my habit has become worse every year since.  Now, most every aspect of my life is connected.

The good news is that I’m not alone. According to Forrester the number of Internet users will grow from roughly 900 million in 2008 to 3.6 billion in 2016 – a 4x increase. Furthermore, according to OpenDNS the number of web queries (a good proxy for activity) will grow from 3 billion to over 310 billion over the same time period, a 100x increase.  So not only are there 4x as many users, but they will be 25x more active than they were just a few years ago.

The “Internet” was not architected for this level of traffic and as a result we increasingly see outages, inflexibility, soaring management and capital equipment costs.  To support billions more, more actively engaged users will require a re-think of digital infrastructure.

This digital infrastructure, essentially the building blocks of the Internet, consists of two parts – Carrier Networks and Data Centers.  Carrier Networks are the ways we connect to the Internet [ mobile, Wi-Fi, broadband] and the Internet backbone.  According to Cisco, mobile data traffic will grow 26x over the next 5 years as more applications and services get served via the cloud.  Is it realistic to think that Verizon, Comcast or Telefonica is going to buy 26 new pieces of infrastructure for every one they have today?  No, they will need to find new ways to solve the problem.

Digital infrastructure also includes Data Centers run by large corporations, web companies and cloud providers.  Data centers are rooms filled with computers, storage, networking and databases that deliver applications and services. Cisco forecasts data center traffic to grow 4x over the next 5 years. 

But it’s not only the volume of activity, it’s the type of activity.  In the old days, traffic in a data center typically flowed “north-south”.  That is, a user request would come into the data center network, get directed to the appropriate server and that server would send the response back to the user.  Nowadays, “applications” are distributed across dozens, hundreds, even thousands of different servers that all work together to fulfill a request. The amount of traffic between these servers (to prepare the response) is approaching 6x the amount of traffic of the response itself.  For example, my Facebook homepage has 4MB of text, images and links.  It’s likely that this page requires nearly 24MB of “east-west” traffic within servers on Facebook’s network to assemble the page – and this ratio is only growing as apps get more distributed.

As a result, a networking perfect storm is brewing – huge increases in traffic, changes in the nature of traffic and a dramatic increase in the size of data centers (now as large as football fields, spurred by major cloud and online services providers such as Amazon, Google, IBM, and Microsoft). In fact, according to IDC, total data center space will increase significantly, growing from 611.4 million square feet in 2012 to more than 700 million square feet in 2016 fueled by the rise of these mega data centers.

One of the great things about being involved in the early-stage investment business is it gives you a front row seat to cutting-edge shifts in technology. We get to do meaningful work -- partnering with visionary entrepreneurs who believe they can build a company that makes an enormous impact on the world. Oftentimes though, their vision describes a completely new way of doing something that challenges accepted conventions – and this can be difficult to grok. 

Here at North Bridge, we’ve known Ash Ashutosh, CEO of Actifio, for over a decade. In 2009, he set out with a vision for a “radically simple” copy data management solution to solve a problem that’s only getting gnarlier as enterprises gather, analyze, and store data at a galloping 45 percent CAGR (IDC). Up to 75 percent of a company’s data is simply copies of primary data that have been created in various non-integrated, siloed processes – backup, snapshot, business continuity, disaster recovery, testing & development, and analytics. Amazingly, many enterprises are now paying more to manage copies of their data than to manage the original data that actually creates value. Enterprises can’t continue to carry the cost and complexity associated supporting all these multiple copies – especially as Big Data keeps getting bigger. No one benefits other than the old storage guard.

This is the $34B+ global market opportunity waiting for someone with the vision and IP to address it. Solving these types of problems is hard and many are uncomfortable wading into uncharted waters where there is no proven model. Based on our prior experience with Ash, we knew how smart he was, how deeply he understood customer problems, of his unique insights into the storage market and that he’s a tenacious entrepreneur. With this trust and our sector expertise we jumped in to brainstorm, validate the ideas with customers and build the company together. 

Along the long road from inception to impact are a number of milestones that indicate the hard work, the beliefs and the loyalties that North Bridge shares with its founding teams are beginning to make a difference. Today's announcement from Actifio that North Bridge is joined by some of the world’s leading VC firms in a $50 million investment in Actifio, is one of those milestones – and validates that the Ash’s original vision has gained great traction.

The fact is attacking the copy data problem is knotty and complex, especially when the goal is to make the solution simple. Yet, Actifio and its collection of industry vets have engineered a solution that saves a fast-growing roster of happy customers up to 90 percent in copy data management costs. This is a great example of North Bridge’s bias toward startups that attack big, hairy problems in far-sighted ways with the potential to transform entire industries. Because Actifio is based in Waltham, Mass., it’s able to call on some of the best infrastructure engineers in the world, right here in New England. Actifio was Storage Magazine’s 2012 Product of the Year, and last month was named one of Forbes’ Most Promising American Companies, which validates our belief that to be in the right place at the right time, you go to the right place and work hard.