VC firm Andreessen Horowitz recently announced their fifth fund. In a recent podcast, the firm's founders made several predictions that could impact the enterprise data center.
Venture capital firm Andreessen Horowitz (A16Z) has come to be known by Marc Andreessen's famous tagline that "software is eating the world." The premise is that innovation happens at the logical layer. While the innovation is packaged in hardware, the change happens via software. Looking at both cloud computing and software-defined infrastructure, it's easy to see how software is impacting the data center.
A16Z has an attractive portfolio of investments. Beyond the incredibly popular consumer-facing brands such as Facebook, Groupon, Lyft, and Foursquare, A16Z has some notable enterprise technology investments as well. Their industry-focused investments include Actifio, Barefoot, Box, Cumulus, and Nicira to name a few.
In a recent podcast, founders Marc Andreessen and Ben Horowitz discussed their vision for the firm's recently-announced fifth fund. The discussion prompted several thoughts around how their predictions will impact the data center. Here are three top takeaways from their predictions for IT professionals.
1. Processor pricing
The pair's boldest prediction is that Moore's Law for increasing processor performance approximately every two years is now inverted. Processing power becomes cheaper opposed to processors becoming faster. While CPU power remains stagnant, the cost of the same CPU power is halved. A16Z believes the trend will continue until processing power is effectively free to acquire.
Free processing has obvious implications from an enterprise data center perspective. For example, if the cost of computing power halves every two years, what is the impact on the cost to benefit ratio of public cloud computing? The concept of virtually free processing power leads to the foundation of many of the follow-on predictions. These follow-on predictions are some of the most interesting impacts on the enterprise data center.
2. Cloud computing
Remember the premise of A16Z's investment strategy is that software is eating the world. Cheaper, and therefore abundant, processing power isn't innovation in itself. The availability of increased computing capability enables software-driven innovation. The cost of infrastructure related services is reduced. As such, aggregated services from companies such as Amazon, Google, and Microsoft could be reduced as well.
The cost reduction and increase in capacity for cloud providers creates two tracks of opportunity. The first is the ability for cloud computing companies to offer additional services. Look no further than machine learning. The availability of better algorithms powered by cheap CPUs and GPUs enables service providers to create profitable business models behind artificial intelligence (AI) as a service.
The second track is innovation based on the abundance of AI. Additional industries will experience disruption as innovation springs up, built on the availability of cheap AI as a service. Internet of Things (IoT) devices armed with ever-increasing CPU power create abundant sources of machine learning. The idea of intellectual property (IP) shifts, too. Enterprise data center managers need to understand the impact of the shift in what's valuable IP versus what's commodity data, and adjust security policies accordingly.
3. Every device networked
Another prediction is that every CPU will connect to the network. This prediction has tangible impacts for the data center manager as well. Specific questions to ask are: How these devices are attached to the network, what are the security implications of having exponential network growth, and how are performance planning and capacity planning impacted?
I foresee some tough architectural questions around networking for large enterprises as a result. All of this innovation requires integration with legacy infrastructure. Will network managers be better served by replicating enterprise network designs across multiple cloud vendors? A single logical network domain is the promise of solutions such as VMware NSX. Or, will cloud-native applications take hold and new paradigms in network management form to integrate cloud and legacy architectures?
What do you think?
I'd love to hear your thoughts on the subject. Is there a massive shift approaching as indicated by some of the A16Z predictions? Or, is it business as usual for the next several years? Share your opinions in the comments section or on Twitter.
- How to tell if a GPU-oriented database is a good fit for your big data project (TechRepublic)
- IoT developers: Master this coding language if you want to thrive (TechRepublic)
- Why the future of a Dell-led VMware hinges on developers (TechRepublic)
- Intel slows the rate of major chip upgrades as Moore's Law falters (ZDNet)
- Why open source will be critical to the future of SDDC (ZDNet)