[email protected] Consistent and Comprehensive Hybrid Cloud Azure Services Everywhere Microsoft Global Dark Fiber Investments
Total Page:16
File Type:pdf, Size:1020Kb
[email protected] Consistent and comprehensive hybrid cloud Azure services everywhere Microsoft global dark fiber investments DCs and Network sites not exhaustive Owned Capacity Data center Leased Capacity Svalbard Future Capacity Edge Site Greenland United States Sweden Russia Norway Canada United Kingdom Poland Ukraine Kazakistan Russia France Turkey United States Iran China Algeria Pacific Ocean Atlanta Ocean Saudi Libya Egypt Mexico Arabia India Myanmar (Burma) Niger Mali Chad Sudan Pacific Ocean Nigeria Venezuela Ethiopia Colombia Dr Congo Indonesia Peru Angola Brazil Zambia Indian Ocean Bolivia Nambia Australia South Africa Argentina Microsoft Confidential 4 disconnected Intelligent Cloud & Intelligent Edge Examples Sensors + Control Sensors to Interactive Integrated Platform Global scale processing Microcontroller IoT Devices Edge Devices Edge Appliances Hybrid Cloud Hyperscale Azure Sphere Azure IoT Device Azure IoT Edge Azure Data Box Azure Stack Cloud SDK Edge Azure Regions • Highly-secured, connected • Deploy and manage Azure • Cloud Consistent Edge MCU • 1000+ devices Services in containers on • Data Box Edge: AI- • Edge and Disconnected • Full Range Hyperscale any IoT device • Azure Sphere Linux OS for • 250+ partners Enabled, Storage and Scenarios Cloud Services modern MCUs • AI, AzureML, Azure compute Azure Edge • All certified to work great • Regulatory Requirements • Tiered Service availability: • Azure Sphere Public Stream Analytics and appliance Heroes > Hubs > Satellites with Azure IoT Hub more • Cloud app model on- Preview release in • Includes Azure IoT Edge premises • Open Source Based September 2018 • Generally available Services & Tools • Included Azure IoT Device • Data Box: Offline, SDK ruggedized data transport, 100 TB – 1 PB Inferencing Inferencing Inferencing Inferencing Inferencing Inferencing Training Training Training Training Training Training Data Data Data Data Data Data Custom AI model Custom AI model Custom AI model Custom AI model Custom AI model Custom AI model Less aggregated data and insights Full spectrum of data collection and aggregation using Azure Storage More aggregated data and insights Less AI functionality Azure AI functionality availability based on computational resources Most AI functionality Intelligent Cloud & Intelligent Edge Examples Sensors + Control Sensors to Interactive Integrated Platform Global scale processing Microcontroller IoT Devices Edge Devices Edge Appliances Hybrid Cloud Hyperscale Azure Sphere Azure IoT Device Azure IoT Edge Azure Data Box Azure Stack Cloud SDK Edge Azure Regions Mediatek • Full Range Hyperscale Cloud Services • Tiered Service availability: Heroes > Hubs > Satellites AI-Link • Open Source Based Services & Tools Microsoft Azure Databox Edge Qualcomm Vision AI Dev Kit USI Seed Microsoft Azure Stack - Avanade - Cisco Intel Azure IoT - Dell Edge certified kits - Fujitsu - HPE - Lenovo - Wortmann Intelligent Cloud & Intelligent Edge Examples Sensors + Control Sensors to Interactive Integrated Platform Global scale processing Microcontroller IoT Devices Edge Devices Edge Appliances Hybrid Cloud Hyperscale Azure Sphere Azure IoT Device Azure IoT Edge Azure Data Box Azure Stack Cloud SDK Edge Azure Regions Mediatek • Full Range Hyperscale Cloud Services • Tiered Service availability: Heroes > Hubs > Satellites AI-Link • Open Source Based Services & Tools Microsoft Azure Databox Edge Qualcomm Vision AI Dev Kit USI Seed Microsoft Azure Stack - Avanade - Cisco Intel Azure IoT - Dell Edge certified kits - Fujitsu - HPE - Lenovo - Wortmann Azure Azure Azure Stack Azure DevOps Azure Public Clouds Visual Studio Azure Repos Azure Pipelines Azure Pipelines Developers Continuous Integration Continuous Deployment Visual Studio Code Azure Stack Clouds Border Devices Leaf Switch Leaf Switch BMC Switch Hardware Lifecycle Host Scale Unit Capacity - Small End Border Devices 64 Cores 6 TB of RAM 384 TB of Storage Leaf Switch Leaf Switch BMC Switch Capacity - Large End 448 Cores (physical) 24 TB of RAM 1 PB of Storage Hardware Lifecycle Host Scale Unit Intelligent Cloud & Intelligent Edge Examples Sensors + Control Sensors to Interactive Integrated Platform Global scale processing Microcontroller IoT Devices Edge Devices Edge Appliances Hybrid Cloud Hyperscale Azure Sphere Azure IoT Device Azure IoT Edge Azure Data Box Azure Stack Cloud SDK Edge Azure Regions Mediatek • Full Range Hyperscale Cloud Services • Tiered Service availability: Heroes > Hubs > Satellites AI-Link • Open Source Based Services & Tools Microsoft Azure Databox Edge Qualcomm Vision AI Dev Kit USI Seed Microsoft Azure Stack - Avanade - Cisco Intel Azure IoT - Dell Edge certified kits - Fujitsu - HPE - Lenovo - Wortmann Replace Image Replace Image Network data transfer from edge to cloud Data pre-processing at the edge Machine Learning inferencing at the edge Easily and quickly transfer data to Azure to Analyze data from on-prem or IoT devices to Run Machine Learning (ML) models at the enable further compute and analytics or for get results quickly and close to where data is edge to get quick results that can be acted on archival purposes. being generated. Transferring the full data set without round tripping to the cloud, while to the cloud or filter or transform it before transferring the full data set to Azure to retrain uploading to save on bandwidth. and improve your models. Azure Storage blob, page, file Storage Gateway Data Box Edge IoT Edge SMB/NFS Shares On Premises Data Aggregate Modify Filter Combine or standardize data from different Remove data that for legal or compliance Sensors and devices can generate huge sources into the format you want for cloud reasons can’t be stored in the cloud, e.g. amounts of data, but often most of that data is storage and archiving as part of your upload Personally Identifying Information. repetitive and uninteresting. Identify the pipeline. important data you want in the cloud for further processing or long-term storage and discard less important data. Process at Edge for immediate results Filter with AI analysis at Edge Remove Sensitive Data at Edge Process images and video as they are Constantly monitor traffic camera feeds to Automatically blur PII data, e.g. faces or license generated for immediate results. Drone video detect collisions or “near collisions” and store plates, from images and video before they are footage can be analyzed in the field, or quality one minute of video around these events for uploaded and archived in Azure, protecting control issues can be identified right at the human analysis and model training. Retrain in against privacy issues if you have a legal factory before the product hits the market. cloud and send updated model to edge. requirement around storing PII in the cloud. • • • • • • • • Data Box Edge Azure Cloud Prevent equipment failures by identifying issues prior to failure and fixing them IoT Hub IoT Edge Modules What problem does this solve? • Reduce lost manufacturing time due to Local broken equipment FPGA + Azure ML Storage • Reduce maintenance overhead with cost- effective preventative maintenance vs Azure ML unpredictable repair costs What technology does this use? Azure Storage • Sensors send data over the local network to Data Box Edge • Containerized application uses a custom machine learning model to identify equipment in need of maintenance • Optional: transfer data to Azure for storage, to retrain and improve your ML model, or for additional analytics Data Box Edge Azure Cloud Improve crosswalk safety by extending crossing time as needed IoT Hub IoT Edge Modules What problem does this solve? • Reduce risk of harm to pedestrians from Local insufficient time to cross FPGA + Azure ML Storage • Reduce risk of vehicle accidents with improved traffic light management Azure ML Local traffic Azure Storage What technology does this use? system • IP cameras send data over the local network to Data Box Edge • Containerized application uses a custom machine learning model to identify pedestrians in need of additional crossing time • Optional: transfer data to Azure for to retrain and improve your ML model or for additional analytics Data Box Edge Azure Cloud Utilize and store high-value business data while remaining compliant with regulations IoT Hub IoT Edge Modules What problem does this solve? • Mitigate risk of collecting, using, and Local storing many types of high-value FPGA + Azure ML Storage Cognitive Services business data What technology does this use? • Various sensors or on-premises systems send On Premises Systems data over the local network to Data Box Edge Azure Storage • Containerized application uses Azure Cognitive Services to detect and remove objects from images or text • Optional: transfer data to Azure for storage or additional analytics Data Box Edge Azure Cloud IoT Hub IoT Edge Modules Local Storage FPGA + Azure ML Azure Services On Premises Systems Azure Storage Offline Data Transfer Online Data Transfer Data Box Data Box Disk Data Box Heavy PREVIEW Data Box Gateway Data Box Edge • Capacity: 100 TB • Capacity: 8TB ea.; 40TB/order • Capacity: 1 PB • Virtual device provisioned in • Local Cache Capacity: ~12 TB • Weight: ~50 lbs. • Secure, ruggedized USB • Weight 500+ lbs. your hypervisor • Includes Data Box Gateway • Secure, ruggedized drives orderable in packs of • Secure, ruggedized • Supports storage gateway, and Azure IoT Edge. appliance 5 (up to 40TB). appliance SMB, NFS, Azure blob, files • Preview: September 2018 • GA September 2018 • Currently