Editor’s note: Jillian Mirandi is Senior Analyst at Technology Business Research.
HAMPTON, N.H. – Amazon’s core online retail business spawns yet another high growth off-shoot with Machine Learning
On Thursday, Amazon Web Services (AWS) Senior VP Andy Jassy announced the release of Amazon Machine Learning, a service that allows developers and enterprises to more easily integrate data-driven insight into applications. This announcement holds tremendous upside, connecting AWS at the core of emerging trends like big data, internet of things, and advanced analytics.
The service draws directly from existing strengths within the Amazon business model. Just as AWS itself was formed from Amazon’s IT proficiency and efficiency, so too is Machine Learning born from the company’s process for analyzing and acting on the reams of data being collected from its own customers.
Furthermore, not only has Amazon proven its analytics prowess with the retail recommendation engine, by providing the back-end computing for customers like NetFlix, it already has a track record of supporting web-scale applications with embedded analytics functions for customers.
TBR expects the machine learning service to become yet another significant growth driver for AWS’ cloud business. The announcement is well timed, as AWS is both facing more competition in the public cloud space and difficulties connecting its public cloud services to growing demand for hybrid clouds that combine both on- and off-premise IT infrastructure. AWS brings both the largest base of public cloud customers and developers to bear in this effort, which will provide a large testing ground and target base for the newly announced service.
Though more vertically-oriented IoT providers will deliver much more tightly packaged solutions, we expect AWS to be a significant presence in big data analytics and machine learning for cloud services with this introduction.
Simpler analytics is the latest effort to expand AWS’ developer relationships
The role of developers in AWS’ overall go-to-market strategy is increasing. On the same day of the Machine Learning announcement, AWS also touted a number of large software vendors, including MicroStrategy, Software AG, and TIBCO, that have standardized on AWS for cloud service delivery. These relationships drive AWS usage and provide a very stable customer base for the underlying services. To further that developer-focused initiative, Amazon has recognized that developers need help organizing data and building models from complex data sets. Two of the biggest barriers for developers to integrate insight into their applications are structuring the data and auditing the quality of that data. AWS has built features to address both these issues into the machine learning service. Amazon Machine Learning will use a company’s existing data for data visualization and exploration, sourcing new models.
The product will feature programmatic access via APIs, regression models, binary predictions, multi-class predictions, and machine learning algorithms. Amazon Machine Learning will then take users through data validation and optimization options, helping them make real-time and batch predictions. To further support the developer ecosystem, Amazon notes that mobile iOS and Android will also be updated to support prediction requests.
Pricing is pay as-you-go, set at $0.42 per compute hour for data analysis, model training and model evaluation. Batch and real-time predictions cost $0.10 per thousand; real-time predictions are also charged an hourly reserved capacity fee.
(C) TBR
