Editor’s Note: Robin Vasan is a Managing Director of Mayfield, a global early-stage venture capital firm. Couchbase, Hashicorp, & MapR are current Mayfield investments.
As we wind down 2014, a wave of innovation is sweeping infrastructure software and providing outsize opportunities for startups.
From the introduction of Amazon Web Services in 2006 and the beginning of infrastructure-as-a-service, to the unveiling of Github in 2008 and the continued expansion of open source software, infrastructure technologies have evolved at a frenetic pace.
There has been the original launch of Hadoop in 2005 and the rise of NoSQL in 2009; the introduction of the iPhone and Android phones in 2008 as computing became mobile and, more recently, the 2012 acquisition of Nicira by VMWare, public since 2007, which ushered in the new era of software-defined everything.
Consequently, buyers and users of infrastructure software today expect it to be infinite in terms of scalability, resilient in the sense of being always-available and immutable as in true to its original state. Many companies are being formed at all levels of the stack that will grow into giants of the infinite, resilient, immutable software infrastructure (IRIS) era.
Containerization, database domination and the rise of DevOps will all play a part in the infrastructure paradigm, and several startups are emerging as rising contenders in each category.
2014 was the year when Docker emerged as the open source containerization standard, although it seems as if there might be other contenders to the throne. As dockerized applications move into production environments, significant opportunities exist for start-ups to provide persistent and high-performance storage layers, enhanced security and policy management, efficient network interoperability and optimized scheduling and monitoring. These Web scale architectures create opportunities for the new Cisco, Oracle, Microsoft and VMWare to emerge – entrepreneurs, take note.
VMware enabled enterprises to gain significant savings and reduce complexity through server consolidation, but primarily for Windows-based applications.
Today’s enterprises are running a wide array of different operating systems as well as applications built on public and private cloud platforms. Virtual machines and new application packaging technologies (AMIs, containers, etc.) start to provide the foundation for further decoupling software from hardware.
Nutanix is a great example of a converged solution initially built for the virtual desktop use case. All layers of software defined compute, storage and networking are underway, but there are still lots of places to target specific workloads and reduce cost/complexity and improve performance/security.
After being taught for years that all structured data should be put into a relational database or a data warehouse, application developers have been adopting a range of new data platforms.
Hadoop and NoSQL have clearly emerged as preferred approaches for analytical and operational workloads respectively, and companies like Couchbase, Cloudera, HortonWorks, MapR and MongoDB are taking market share from incumbents.
However, there appears to be an early realization that there are a number of common (and perhaps distinct) access patterns for data: relational/sql, document/json, caching/key-value, analytical/columnar, search/inverted, metrics/time-series, files/hdfs and multidimensional/graph. A new wave of data and analytics companies is being formed to exploit these opportunities.
Businesses of all sizes are wrestling with how to leverage IRIS capabilities for a range of their existing solutions. Enterprises are evaluating alternative approaches to either migrate or replace legacy business applications and infrastructure, but they want strong security policies around access and data. An early hybrid cloud use case that was pioneered by StorSimple (acquired by Microsoft Azure) entails archiving data into the cloud via an enterprise storage gateway. A new idea of “bring your own storage” (BYOS) might enable corporations to leverage their existing public or private cloud storage and therefore feel more comfortable about data privacy and controls.
Combining new data platforms and new massively distributed systems technologies will enable companies to store and process billions of data points in near real-time. There are recent successes in business intelligence like Tableau and Qlik, but native analytics and visualization solutions for Hadoop and NoSQL are still yet to emerge.
The application of complex analytics (and real world simulation models) can and will be applied across functional areas like sales, marketing and HR as well as common business processes such as recruiting, budgeting/forecasting and ultimately into specific vertical industry domains like retail, manufacturing and financial services. Palantir is just one example of a successful application leveraging new platforms and massive analytics, but there will be many more.
The Rise Of DevOps
Modern applications are built from lots of components (open source, APIs and other services) and being run on top of private and public cloud environments. These new complex distributed systems are increasingly complicated to secure, orchestrate, manage and monitor. Google, Amazon, Facebook et al have armies of engineers, but they have also invested in technology to help them scale to literally millions of machines.
The Road Ahead
2015 will present an area of great opportunity for entrepreneurs founding infrastructure and application software companies. By remembering the elements of IRIS and targeting any of these six areas, startups can grow into lasting companies.