The 2018 Microsoft Ignite conference was overflowing with attendees this year, as user enthusiasm continues to grow with the advent of CEO Satya Nadella. Since space is short and the announcements are many, let’s get straight to the details of the product innovations available for data professions.
New in AI
If you’re interested in voice and natural language processing, then here are a couple of announcements just for you:
- Not long ago, Microsoft had several distinct artificial intelligence (AI) speech capabilities, including speech recognition, speech translation, and customized models to create a unique voice for your app. Those have been combined into Speech Service, now generally available. In public preview is Human Parity Text to Speech, which utilizes the latest in deep neural network technology to make computer voices nearly indistinguishable from real humans. Details are at https://bit.ly/2PtWT5r.
- Furthering the goal of making your apps behave even more similar to that of a human being, the Microsoft Bot Framework v4 SDK is now generally available. You can program in C#, Java, Python, and JavaScript. It includes many tools for simplifying and building bots with a modular and extensible architecture, allowing easy selection of the components and services you need. Details are at https://bit.ly/2OSpJwR.
- Want your apps to include a conversational personal intelligent assistant (PIA) that knows how to complete specific corporate tasks? Check out the new Cortana Skills Kit for Enterprise. Underpinned by Azure Bot Services and Azure Language Understanding, imagine scenarios such as building a Cortana skill that lets employees ask the PIA to schedule their vacation time, saving 30 to 60 minutes of intranet surfing to figure out how to make the arrangements themselves. This program is by invitation only and not all features have been announced. See https://bit.ly/2ArAIEz for more information.
My personal interests in AI mostly focus on machine learning. Here, Microsoft further advances its machine learning capabilities and also puts more AI into its own products, such as in Microsoft 365, Excel, and Dynamix 365.
- New in Azure Machine Learning is the automated machine learning feature set which enables users to automate: data transformations at speed, model selection for more efficient algorithms, and hyperparameter tuning to quickly learn the accuracy of a given pipeline’s predictions. The new SDK for Python includes features for distributed deep learning, allowing massive clusters of GPUs, access to field programmable gate arrays (FPGA) for amazing speed at image processing, and easier integration with IDEs such as Visual Studio Code, Jupyter notebooks, Azure Databricks notebooks, and PyCharm. Read more at https://bit.ly/2PSp2Qp.
New in the Data Platform
Microsoft used Ignite 2018 to announce the public preview of SQL Server 2019, with a special focus on big data features. First, Spark and Hadoop distributed file system (HDFS) are now built into SQL Server to accelerate ingestion, storage, and analysis of data at petabyte scale. There are also new direct query connectors to Oracle, Teradata, and MongoDB. As with every release, there are many security, usability, and optimization improvements. But I will cover those in a separate, more detailed article.
Details are at https://bit.ly/2CFG8x1.
Azure SQL Database offers improved query performance features under the Intelligent Query Processing moniker. This feature set includes row-mode memory grant feedback, approximate query processing, and table variable deferred compilation. All three features work together to make SQL processing faster, more responsive to memory usage issues, and improve long-standing SQL coding issues. And, as happens with every release, Microsoft has announced more pricing and performance tiers for great flexibility (and potentially lower cost) when deploying apps to the cloud, such as the new Azure SQL Database Hyperscale, a way to deploy a single but highly scalable database that can grow from a few to hundreds of terabytes in size. Azure SQL Data Warehouse also provides a new, lower-entry point to help customers get started more quickly. Details are at https://bit.ly/2NKNOVw.
Microsoft’s NoSQL offerings have also grown:
- The Spark-like Azure Databricks product now provides Azure Databricks Delta as public preview, improving data reliability, simplified data pipelines, and improved job and query performance. More information is at https://bit.ly/2NRBtz5.
- Azure Cosmos DB provides multi-master support for high availability with single millisecond latency and better conflict resolution. The Reserve Capacity feature reduces costs for using Cosmos DB. And the Cassandra API is further enhanced for users familiar with that NoSQL platform. Details are at https://bit.ly/2RgbMoR.
- Azure Data Explorer is in public preview. This is a new indexing and querying service for interacting, lightning-fast ad hoc data exploration from data that originates in apps, servers, and edge devices. Details are at https://bit.ly/2yuQPzN.
Other News for Data Professionals
There’s more than one way to learn what’s new, starting with the new Microsoft Learn and the role-based Microsoft Certifications Microsoft Learn. This is free, interactive web-based training with a tutorial approach to teach Azure and Business Applications. There is cool progress tracking and gamification features such as achievements, while using free Azure resources for hands-on learning. There’s also a new role-based set of Microsoft Certifications aligned to job roles.
Go to https://bit.ly/2N8hGWL to learn more.