Analysts predict that there will be anywhere from 24 to 200 billion Internet connected devices by 2020. Imagine every device as an object that is capable of exchanging voluntary information about itself and its environment as well as executing commands via APIs to act on their environments. What's more, the analytical insights gleaned from this data will have the potential to transform existing business while creating new ones. The Internet of Things is here, and the heterogeneity of devices, messaging considerations, security constraints, volume of data and compute requirements are significant. In this talk, seven-time MVP and Microsoft Azure MVP Rick G. Garibay will provide a deep dive into the various ways in which devices can be connected today and how a service-assisted communication model is really key to tackling the myriad of constraints that exist today. We'll explore the strengths and weaknesses of various communication models and application protocols and how the Azure Reykjavik Device Gateway Accelerator addresses your needs today while providing the extensibility and scale to meet your needs as your IoT strategy grows. You'll learn how to get up and running on your Azure subscription with the Master and Partition roles, how to register a device and flow telemetry, inquires, commands and notifications through the Reykjavik fabric leveraging Azure Storage, HD Insight along with a look at how you can extend the fabric with custom protocol gateway and telemetry adapters.
Any sample code, video, or other supporting content for this presentation (if available for this session) is loaded in the zip file.
You've migrated parts of your application to in-memory. But now you are not seeing a serious performance improvement, or maybe even a performance degradation! What's going on? In this session we walk you through several performance problems we have seen with customers, and show how to diagnose and resolve the issues to ultimately obtain the performance improvement you're looking for from in-memory OLTP. We address a variety of issues, including log IO bottlenecks, index tuning, and query plan problems.