Editor’s note: Navin Chaddha is managing director of Mayfield, an early-stage venture capital firm. He holds 35 patents, is a serial entrepreneur, has been involved with two dozen IPOs and acquisitions, and is a lifelong cricket fan.
The software-defined data center should be a call to all entrepreneurs focused on enterprise infrastructure that the past does not predict the future, but it can be undone by it.
And the past was great. It was great for entrepreneurs and VCs. Categories for hardware-based servers, storage, networking and security were created in the past 30 years that still generate $200 billion-plus in annual sales. The margins were fat, product lock-in was substantial and barriers to entry were high. The game plan worked for everyone and everyone had the same game plan: Focus on hardware.
But that game plan no longer works. Established infrastructure categories are getting disrupted with newer technologies that disaggregate software from hardware, all of which has now been reduced to commodity boxes stripped down to their essential components and nothing more. The innovation now starts (and ends) with software.
Today’s enterprise infrastructure startup leaders need to be focused on pairing innovative software with commodity off-the-shelf (COTS) hardware. They are, in essence, software “infrapreneurs.”
The software infrapreneurs we’re seeing are coming out of either Internet-generation companies like Google, Facebook, Amazon and Yahoo, or had significant terms at hardware companies such as Cisco, EMC, Hewlett-Packard and Sun Microsystems/Oracle.
To individuals in both groups, the advantages of software-led companies are obvious. They recognize that a hardware-focused approach has limitations and that the real opportunity in infrastructure today is via software (see also Marc Andreessen’s article in The Wall Street Journal from 2011 on “Why Software is Eating the World” for software’s impact on a number of different industries).
These CEOs are passionate about design, conscious of the power of community, are willing to sacrifice tradition, and bring different business models, sales strategies and license models vs. past leaders. The changes they will create are predictable — more flexible technology, easier to use, based on open technologies, and cheaper to use — but the winners and losers are still being decided.
The server market is a good example of what will happen to other hardware infrastructure segments. VMware figured out how to separate software from the server it runs on, increasing the operational and economic efficiency of data centers everywhere. Virtualization technology spread, sucking up the innovation previously contained in the servers, reducing the leverage that hardware vendors had on their customers.
As software did more, hardware did less. Server economics are now defined by high volumes and low margins. The changes here have played a part in deep structural overhauls at the world’s three largest server manufactures: Dell went private in 2013; IBM exited the market last year by selling its server business to Lenovo; and Hewlett-Packard is in the process of splitting itself into two. Through it all, VMware grew stronger.
So, while the server segment of the data-center ecosystem is solved, other areas remain unsolved. Altogether, enterprise infrastructure represents a more than $200 billion opportunity. This isn’t a comprehensive list, but this is what we see as the biggest opportunities for infrapreneurs looking to impact the future:
Compute: We mentioned VMware’s undoing of the server market. Now it’s faced with its own disruptive threat in the form of container technology from the likes of Docker and CoreOS. The infrapreneurs we’re seeing here are changing how apps are developed, deployed and managed. This is an area that is relatively new but has taken hold at the most fundamental levels of the data center. We’re now seeing advancement here from infrapreneurs on various fronts.
Networking: Networking is undergoing tremendous upheaval. This hardware appliance-dominated market is being ripped apart by software-defined vendors like Cumulus Networks’ Linux OS for switching and Versa Networks’* network function virtualization (NFV) for L4-L7 services. New companies in this segment are riding the COTS hardware wave and providing capex and opex savings, networking automation, consolidation of different network functions, agility and scalability.
Storage: Software is invading proprietary storage systems from multiple angles, upending the business of legacy hardware storage vendors. InkTank (acquired by RedHat) and Swiftstack* are examples of companies providing software-defined storage on COTS hardware. These two, among others, are helping to reduce capex and opex significantly and enable scale-out architectures. Additionally, disruptors in this segment enable the purchase of heterogeneous storage hardware without having to worry about 1) interoperability issues, 2) under- or over-utilization of specific storage resources and 3) manual oversight of storage resources.
Security: In the security space, there are two primary disruptions happening: On-premise security products are being remade into cloud products and a new generation of companies are providing security functions in software on COTS hardware. New age companies that provide software on COTS hardware are moving beyond protecting traffic at the perimeter of a data center to protect data as it moves laterally within a software-defined data center and can create the next Palo Alto Networks. Demand for this sort of technology is a direct result of new capabilities required by an increasingly software-centric world.
Management: We also see a need for next-gen systems management companies that can manage this new world. What will it take to build tomorrow’s IBM Tivoli? The Holy Grail here is a single dashboard that can represent the health and status of all of an enterprise’s systems. We haven’t seen it yet, but we’ve been around long enough to know that when a great problem exists, a great solution is in the works.
For each of the above, the common linkage is software. The opportunity throughout the software-defined datacenter space is to implement software-centric subscription sales models to disrupt the business model of legacy vendors selling via upfront purchases of hardware appliances.
Infrapreneurs will find plenty of opportunity in the enterprise infrastructure space for years to come if they stay focused on using software as the means to innovate. The future is waiting.