Editor’s note: Chris Potter is co-founder of ScreenLight, a cloud video collaboration service for video producers to review and approve video with their teams and clients. Follow him on Twitter @potta.
Google recently rolled out a number of enhancements to its Cloud Platform products. With these changes, it’s clear that Google is aiming its sights directly at Amazon Web Services, not to mention other infrastructure-as-a-service providers, such as Windows Azure, Rackspace, HP, and other OpenStack providers.
With its global reach, Google has the economies of scale that could eventually allow it to underprice all but Amazon for raw computing power, network connectivity, and other basic infrastructure services. However, going head to head with Amazon has significant challenges, not the least of which is Amazon’s significant head start in developing a rapidly increasing number of very granular services, as well as its success in winning the hearts and minds of developers.
While it’s still early days in this market, if Google is going to displace a strong and entrenched competitor such as AWS, it’s going to have to be incrementally better or significantly cheaper. Alternatively, Google could try to redefine the game to be one that it has a greater chance of winning.
Rather than competing head-on, perhaps a better strategy for Google would be to build on its strength of developing highly scalable web and mobile applications. While all of its competitors focus on selling infrastructure as a service, Google could pair that with pay-per-use API services that could be linked together by developers to build upon.
A focused set of services that solve common problems or eliminate time-consuming development and integration tasks would help developers build apps faster. This would be of use to both application developers and enterprise customers looking to move to the cloud. Some examples of these services, which will be examined in more detail below, include accelerated file uploads, video encoding, and real-time communication services.
There is no doubt that there is a market for these types of services. Well-designed API-based services allow developers to move more quickly and focus on the unique differentiators of their applications. Right now, a challenge for developers is that they have to select and test the APIs of a number of different service providers before deciding which ones to integrate. If each one has a different pricing model, terms of service, and documentation style, this adds transaction costs and negates some of the savings. As such, developers could benefit from getting some of these core APIs from a single provider under a single business model and set of terms. Google API services could provide a secure, reliable, and trusted foundation for development.
Platform as a service (PaaS) vendors like Heroku, Engine Yard, and Joyent have started to fill this role by partnering with a number of companies that provide add-on services that make it easier for developers to build full-featured apps. Heroku alone lists over 100 partners (a number of which offer competing services) that include databases, platform/stack, management and monitoring, and app development APIs for billing, real-time push communication, etc. That being said, they act more like aggregators than a single point of contact.
Google has unique strengths when it comes to building scalable cross-platform web and mobile services. None of the other IaaS or PaaS providers, including Amazon, have demonstrated true capabilities in this area. Google has an opportunity to shift the dimensions of competition and attract developers and enterprise customers to its fledgeling cloud services platform by providing a number of APIs that help them solve some of the core problems associated with building web applications. The following sections detail some of the areas that Google should consider.
Accelerated Data Transfer
One of the bottlenecks with big data applications, video services, and data archiving is transferring massive amounts of data into the cloud. TCP-based transport protocols like FTP and HTTP are inefficient for transferring really large data sets at high speeds.
Amazon has sidestepped this problem by promoting Aspera and its faspex protocol to accelerate uploads. But Aspera is a high-priced partner that is just figuring out how to sell its software as a cloud service. Unpublished pricing, monthly minimums, and an enterprise sales model pose barriers to developers who might otherwise find the services valuable.
Google could build accelerated data transfer into its cloud storage and compute services. While other competitors may have issues developing an accelerated transfer protocol, with projects like SPDY (Google’s proposal to speed up web communications), Google has the talent to solve this problem for its customers.
If Google didn’t want to build this itself, Signiant, FileCatalyst or even Aspera would be acquisition candidates that could speed time to market. Acquiring Aspera in particular would be a thorn in Amazon’s side.
If an accelerated transfer protocol isn’t the answer for Google, there are other ways that it could make it easier for customers to get data into the cloud. One option would be to offer an upload API that reduces the distance that data is transferred by routing upload requests to the nearest point of presence (think of this like the reverse of a content delivery network) and then pushing them to the storage or compute region that the customer chooses via Google’s own fibre network. Filepicker.io would be a good foundation for a Google upload API.
Developers, enterprise customers, life sciences researchers, big pharma, or anyone else with massive datasets would benefit greatly from an easier and faster way to get data into the cloud.
Developers of video applications would benefit from a low-cost video encoding service that is based on Google’s Cloud. Likewise, as the market for video explodes, broadcasters, studios, and over the top video services like Netflix are shifting from on-premise encoding to cloud services so that they can easily convert their libraries into the myriad of codecs and bit rates required to deliver video to growing array of devices.
Others have already stepped into this rapidly growing market, but to the best of my knowledge, none of them operate in Google’s cloud. Encoding.com, which just raised another $1.5 million from its existing investors, uses AWS and Rackspace. Zencoder, which Brightcove recently acquired for $27 million, runs on Amazon’s cloud. The rapid growth of these companies shows that there is value to solving this problem. Microsoft also senses the opportunity in this area and is previewing Azure Media Services that can be used to store, encrypt, encode, manage, and stream media content.
Google could easily leverage the encoding infrastructure used by YouTube, which is the largest encoding system on the planet. Doing so would give it economies of scale that couldn’t be matched by the other cloud encoding providers, and could significantly reduce the cost of encoding adaptive bit-rate video for streaming to desktops and mobile devices.
Another benefit of offering a cloud encoding service is that it would give Google another way to push adoption of WebM if it’s still inclined. Another decision point for content distributors is approaching with the release of H.265 and DASH video streaming. Google could also use its encoding service to influence which standards its customers use.
Adding scaleable real-time communication features to web and mobile applications is another significant pain point for developers. Video conferencing, VoIP, SMS, and chat all use an alphabet soup of protocols.
Developers are currently turning to a number of standalone API services to solve this problem. Twilio, is broadly used for telephony services, while PubNub and Pusher provide increasingly popular APIs that are used to send real-time notifications and messages.
What unique capabilities would Google bring to the table? Well, Google+ Hangouts are a pretty fine way to chat with people and host virtual meetings. Opening up the technology behind Hangouts to developers as an API that they can easily integrate into their apps could unleash a storm of creativity. Offering this technology would also benefit Google by luring people away from Skype and Apple’s open FaceTime protocol. In the war of standards, it’s the technology that is most broadly adopted that generally wins.
In an increasingly mobile-dominated landscape, voice is going to be an increasingly popular way to interact with devices. While Siri may have popularized it, the voice war is far from over. Why not offer this as a technology that anyone can add to a mobile application? The more people use Google’s voice search technology in their apps, the easier it will be for the company to maintain its search dominance (and presumably advertising) on mobile devices.
The Potential For Google’s Long API List
The list of APIs that Google could open up and sell to developers on a pay-per-use basis is pretty long. It already offers some of these: email delivery and image processing through Google App Engine; Google Docs Viewer is already being used by developers (whether the terms of service allow it or not); and others APIs like mapping and location services use standardized terms and can be tightly integrated and distributed through its cloud platform.
Google doesn’t have to offer all of the services itself. For example, it could provide a task queuing service like Iron Worker. In this way, users wouldn’t have to spin servers up and down to scale an app. They could simply pass jobs to Google’s army of workers, and be charged for actual processing time for the tasks they send to the workers. Google could offer a library of its own workers, and it could provide developers a platform where they could upload and share their own worker recipes.
Choosing A Different Path To Unseat AWS
Google has a real opportunity to create a unique and valuable platform that leverages its global connectivity and its infrastructure, and plays to its strengths building massively scalable web and mobile applications. It”s a win-win proposition for app developers and Google.
Developers benefit from a powerful selection of APIs that could be integrated into web and mobile applications. This would help speed time to market and free up development resources that could then be devoted to working on unlocking the unique sources of value of their apps.
For Google, this strategy would have the immediate benefit of attracting developers to its new platform and would provide a tier of services that would differentiate it from Amazon and other pure Infrastructure as a Service providers. As Amazon broadens the scope of its granular services to include data warehousing, with its Redshift data warehouse service, Google can help developers by making it faster and easier to build applications using its services.
This strategy also has deeper platform benefits, as it would spur adoption of standards that Google is promoting in its large-scale battle with Apple, Amazon, and Facebook for dominance of the online landscape.