After being away for a few weeks (vacation and business travel), I returned to the Bay Area and attended GigaOm’s Structure conference and Amazon’s AWS Summit. I have been attending Structure since the first event but this was my first time attending the AWS Summit. I was drawn to these conferences for two reasons. First, while over the past 10 years we have invested and continue to invest heavily in SaaS application companies, we have foregone investment opportunities in companies that provide cloud infrastructure solutions, i.e., PaaS and IaaS. I have been trying to determine whether we should be considering investments in cloud infrastructure companies today, particularly capitalizing on our extensive experience from SaaS. Second, I am always interested to hear about cloud computing best practices that can be used by our portfolio companies.
Cloud adoption is growing among corporations of every size but particularly SMBs. In a report published by Forrester on 4/11 the total size of the public cloud market is pegged at roughly $25B (with the majority today being in SaaS applications) and is projected to grow to $160B by 2020. A recent (May 2011) survey conducted by Morgan Stanley establishes that the usage of public clouds for a variety of workloads is expected to show a 23% CAGR over at least the next three years, with higher usage rates for SaaS applications compared to cloud-based infrastructure. SMBs see the use of public clouds as a means of significantly reducing their hardware costs. Larger enterprises are using the public clouds more for rapidly extending the functionality of internally developed applications, with mobile extensions cited most frequently, as well as for executing compute-intensive workloads like analytics. There is also a lot more talk about private clouds and hybrid clouds. Most companies in fact are either using all three types of cloud deployment or state their intent to use them. Only 37% of companies surveyed indicated that they will only use public cloud deployments. This means that SaaS application vendors may need to start thinking more seriously about the need to offer versions of their software running on private and hybrid clouds rather than the public cloud options that most, particularly the startups, have today. For investors this will mean that their SaaS companies may need to invest more in R&D and support as they roll out such options.
Security, interoperability, vendor lock-in, reliability, complexity and data privacy continue to be listed as the top inhibitors to cloud adoption. However, ironically customers don’t always feel that public clouds are less secure than their own data centers. But in security they also tend to include regulatory and compliance issues. Privacy is a different story and one that is viewed differently by the US and Europe.
In a report released during the Structure conference, with some numbers corroborated by Werner Vogels, CTO of Amazon.com during his address at the AWS Summit, GigaOm ranks Amazon as the largest public cloud provider with 2010 revenue of $500M with Rackspace being second. In his address Dr. Vogels talked about the investments Amazon has made and continues to make around AWS. Rackspace, Salesforce, Microsoft, Google, IBM, and other large companies that want to be providers of public cloud infrastructure and services, are also making huge investments in order to remain price-competitive and offer the broadest possible set of services. Based on what I heard during that week, and despite the fact that investments in cloud companies are increasing (64 investments in 2009 totaling $180M, and 93 investments in 2010 totaling $713M) I think that it will be hard, if not impossible, for a startup to compete effectively with the large public cloud computing providers. Public cloud infrastructure has therefore emerged as the battle of the giants. I continue to maintain that the SaaS application space offers more opportunity to startup companies and their investors, Trident included.
Cloud continues to be defined by its benefits rather than its technology. Cloud computing’s benefits that were only stated by vendors, are now validated by customers that have been using such solutions. Surveyed customers see cloud computing as a way to lower IT costs, increase corporate agility, and provide the foundation for 21st century architectures with the right levels of abstraction from infrastructure to platform that will result in higher quality systems. Based on these benefits and what was discussed and reported during the conferences, a few of the best practices that may be of use to SaaS application companies include:
- The application’s architecture determines the level of success and the overall cost of deploying the solution to a public cloud. Moreover, the successful use of a public cloud by an application provider requires the continuous collaboration between the cloud provider and the application vendor.
- Public cloud vendors strive to provide uniform, rather than peak-performance, characteristics across the computing resources they provision. This must be taken into account by the SaaS application vendors as they architect their solutions.
- While a public cloud provider could offer a SaaS application vendor with a low cost way to develop and launch a new application, scaling costs continue to be high, often leading the application vendor to create its own public cloud. Of course, companies the size of Zynga and Netflix are able to negotiate special pricing with their public cloud providers, e.g., AWS. However, for private companies that have not yet reached that scale, the options are more limited.
I expect that this year the number of venture investments and the amount invested in cloud computing companies will surpass last year’s numbers. I continue to see investment opportunities the following areas:
- Big Data Analytics. Analysis of Big Data is starting to become synonymous with cloud computing. I expect that this trend will continue as public cloud providers offer Hadoop and MapReduce options allowing analysts to send their large workloads to the cloud. Data movement and “accessibility” of such services by business analysts (you shouldn’t need to be a data scientists just so that you can use such a service) will continue to be issues that will need to be addressed.
- Enterprise mobile applications. Because of the smartphone proliferation, enterprises are now starting to aggressively adopt mobile applications. As a first step enterprises are looking to create mobile versions of their mission critical applications, as well as develop mobile applications in completely new, innovative areas. Public and hybrid clouds are uniquely suited for the rapid development and deployment of these applications. Therefore, both specific enterprise mobile applications and environments for developing and deploying such applications are of interest for potential investments.
- Data and application integration. I have often written about how the proliferation of public and hybrid clouds will increase the need for integrating data to such applications as well as integrating applications. For example, a cloud-based application may need to access data that is behind the firewall, or two cloud-based applications may need to be integrated to effectively automate a business process.
- Data marketplaces and API management. As companies expose more of their data for cloud-based application, the data itself and the APIs through which this access is accomplished will need to be managed. While I think that we are still in very early stages of data marketplaces and API management, and it is not clear whether either of these areas will emerge into substantial markets and lead to the creation of large companies, there may exist a few interesting investment opportunities to consider.
- Accelerating data movement. Moving large data sets from data centers to the cloud remains a vexing problem particularly for big data analytics applications. While cloud-based processing and storage performance are improving, data movement has not yet made corresponding strides and is thus approaches to address this problem are ripe for investment.
I continue to be excited about the prospects offered by cloud computing and delighted by the increasing usage of all its layers (application to infrastructure) by small and large companies. I don’t feel that we have missed big opportunities by not investing below the cloud’s application layer, but feel that there are at least five areas that represent interesting areas for future investments.