2020 might have stopped us from all gathering together in Orlando last week, but it didn't stop the announcements from flowing in. Microsoft Ignite 2020 brought us exciting announcements from across the Microsoft Portfolio and WWT Experts were in attendance (virtually!) to take them all in.
In part one of our key takeaways series, our Azure experts detail some of their favorite announcements.
The new features and capabilities Microsoft brings to Azure every year are amazing. Some are much needed enhancements to existing solutions, like the GA of the next generation Azure VMware Solution (AVS), and others are completely out of this world (we're looking at you, Azure Orbital!). After pouring over the hundreds of exciting announcements, a few trends caught our attention.
- Connect everything — your branches, your satellites… and your refrigerators? Darren Roback, Cloud Platform Architect, explains how Cisco SDWAN + Azure Virtual WAN, Azure Communication Services and Azure Orbital bring it all together.
- Kubernetes everywhere! Microsoft is serious about adoption of Kubernetes, not just on the Azure public, but wherever the workload lives. Casper Pieterse, Cloud Platform Architect, discusses how the Azure Stack Family gets an injection of K8s while ARC Data Services brings the best of Azure Data to your K8s footprint.
- Microsoft wants Azure to be THE platform for your data. Norm Torvik, Practice Lead - Microsoft Data and AI, highlights some of the most exciting announcements on Azure's best data technologies.
Here are a few highlights, directly from the experts.
Connect everything — your branches, your satellites… and your refrigerators?
Darren Roback, Cloud Platform Architect — Microsoft Azure/Modern Work
Cisco SD-WAN Cloud OnRamp integration with Azure Virtual WAN
This was a big announcement last week, as Cisco technology powers a large portion of enterprise network connectivity, and Microsoft Azure's growing popularity as a cloud computing platform. This week at Microsoft Ignite, Cisco and Microsoft announced the public preview of Cisco's SD-WAN Cloud OnRamp for IaaS integration with Azure Virtual WAN.
As a quick recap (you can also find a full deep dive), Azure Virtual WAN is a new Azure networking construct designed to simplify Azure connectivity, routing and security inside of Microsoft Azure. Azure Virtual WAN is built on two foundational constructs: (regional) Virtual WAN Hubs serve as the connection point for all site-to-cloud connections (VPN, ExpressRoute), as well as the interconnection point for Azure Virtual Networks. The (global) Azure Virtual WAN Service then serves to "unite" all deployed Virtual WAN Hubs into a single global network topology.
Microsoft has announced a number of enhancements to the Virtual WAN Service this year, including native support for Azure Firewall/Secured Virtual WAN Hubs, custom routing in Virtual WAN Hubs and (now) support for Cisco SD-WAN NVA appliances in Virtual WAN Hubs.
The Cisco SD-WAN Cloud OnRamp for IaaS integration with Azure Virtual WAN specifically allows for the deployment of Cisco NVA appliances directly into Azure Virtual WAN Hubs, thereby providing end-to-end connectivity and policy declaration from branch offices and corporate headquarters directly into Microsoft Azure. This new integration preserves all Cisco SD-WAN advanced features by guaranteeing all connectivity flows through the Cisco SD-WAN solution.
In this solution, all connectivity from site-to-cloud and within cloud is provisioned directly from Cisco's SD-WAN vManage portal. In addition, Cisco SD-WAN NVA appliances (based on the CSR 1000V platform) can themselves be provisioned into Microsoft Azure directly from the vManage portal.
This integration offers a notable improvement over the previous architecture, which leveraged SD-WAN appliances deployed into a Gateway Virtual Network, and relied on the customer to manage everything behind the CSR appliances. This architecture didn't leverage Azure Virtual WAN, which itself presented challenges for customers seeking to leverage Microsoft's new cloud networking architecture.
As shown below, now the SD-WAN appliances can be deployed directly into Azure Virtual WAN Hubs, providing up to 50Gbps of throughput into Microsoft Azure, while vastly simplifying the overall architecture, manageability, and provisioning of the solution.
What we like about this announcement is twofold — firstly, we're seeing a deeper integration between Azure and the Cisco SD-WAN vManage portal, allowing customers to leverage a single platform (vManage) for the deployment of SD-WAN appliances and integration with Azure networking. Secondly, the native integration of Cisco SD-WAN appliances into Azure Virtual WAN will allow customers to deploy this new networking architecture resulting in a more streamlined deployment of networking resources and site-to-cloud connectivity.
Azure Communication Services
Another big item of news was the public preview announcement of Azure Communication Services, which represents the first fully managed communications platform from a major cloud provider, and joins Twilio as a platform for developers to add real-time communications capabilities to mobile and desktop applications and websites with a few lines of code.
Azure Communication Service is built on top of Microsoft Azure, providing a globally scalable, secure communications platform for developers. The service leverages the communications elements of Microsoft Teams on the backend, while adding a few new capabilities with the service launch (SMS as an example).
As part of the Azure Communication Services launch, Microsoft is providing access to five main communication services:
- Voice and Video Calling Over IP (up to 50 people and including content sharing).
- Chat (one-to-one or group based).
- SMS Text Messaging.
- Telephony Calling (via the PSTN).
- Network Communications Service (WebRTC open standard video calling).
In addition to (now) having the ability to add communications capabilities to applications and websites, developers can also tap into a rich set of Cognitive Services for language translation, sentiment analysis and more. On top of Cognitive Services, developers can also extend the application or website functionality through the use of Microsoft's Bot Framework, Azure Media Services for video broadcasting and access common M365 services (such as calendaring) using Microsoft Graph. End user authentication can also be added, leveraging Azure Active Directory B2C or other popular third-party identity providers such as Google or Facebook.
Microsoft provided an example of where this technology could be useful, detailing an interaction between a customer having an issue with their home refrigerator, with maintenance technicians being unable to dispatch due to COVID-19 restrictions. By adding voice and video communication to the service provider's website, this customer can now interact in real-time with customer service representatives, providing a true omnichannel engagement experience that not only better solves the customer's problem, but also differentiates the service provider in the marketplace.
While this is a rather academic example, it's easy to envision how this could better enable remote healthcare between providers and patients, while also enabling differentiated omnichannel experiences using other popular Azure services. It's because of this that we believe this is a really exciting addition to Azure's service offerings.
This announcement was exciting because, well, who doesn't want to communicate with outer space!
Last week Microsoft announced the launch of a new service, Azure Orbital, which is a satellite ground station as-a-service offering designed to facilitate communication with satellites. With launch prices for satellites dropping significantly in recent years, it is estimated that over 25,000 satellites will be launched within the next five years, adding to the already more than 2,500 satellites currently in orbit. Deploying ground stations is expensive, so Microsoft is launching Orbital to address the market demand and provide a lower cost of entry in this space.
Azure Orbital is a fully managed cloud-based ground station as a service offering from Microsoft, and at launch, Microsoft has committed to a ground station presence in the US (Washington state), Sweden, South Africa, Singapore, Chile and Dubai. In addition to hosting their own ground stations, Microsoft has also partnered with KSAT, Viasat and US Electrodynamics allowing customers to schedule contact with satellites through these providers as well.
Azure Orbital is targeted towards industries such as remote mining, energy farms, defense, remote factories, oil and gas and maritime, allowing operators to facilitate connectivity with these installations using satellite communications. In general, Azure Orbital is targeted towards two key use cases: Earth Observation and Global Communications.
For Earth Observation, Azure Orbital provides self-service scheduling of communications with satellites and facilitates data ingestion and processing directly in Azure. Cloud modems are deployed into the customer's virtual network to demodulate satellite data, and this data can be integrated with a range of Azure services such as Azure Cognitive Services, Azure Storage or Azure Data Warehouse. Customers simply follow a three-step process to get started:
- Register a spacecraft.
- Create a contact profile (including frequency, polarization, etc.).
- Schedule contact with the satellite.
On the Global Communications side, Azure Orbital facilitates high-speed connectivity over satellite networks and through the Azure global backbone. Microsoft has committed to the deployment of teleports in close proximity of Azure data centers or through the interconnection with third party teleports. Through this service, satellite operators can provide Internet breakout at the Azure edge and can integrate value-added cloud services such as SD-WAN, security services, edge computing and even 5G mobile solutions into the environment to provide a differentiated customer experience.
Azure Orbital is now in public preview, and it was mentioned at Ignite that Azure Orbital has already landed their first large customer, with SES selecting Azure Orbital to augment their needs for the O3b mPOWER communication system currently in development.
Casper Pieterse, Cloud Platform Architect - Microsoft Azure
During the 2020 Microsoft Ignite event, Microsoft once again demonstrated that they are embracing the world of microservices and containerization and adopting Kubernetes (and by extension docker) as their go to platform of choice.
Three major Kubernetes announcements caught my attention.
AKS (Azure Kubernetes Service) support for Azure Stack HCI
The Microsoft Azure Kubernetes Services allows administrators to rapidly deploy a fully functional Kubernetes platform using pre-configured, Microsoft provided virtual machine templates that are always up to date and validated. These Kubernetes clusters are automatically hooked up to Microsoft Azure’s monitoring and reporting services allowing administrators to have full visibility of their workloads using the tools and services that they are already familiar with.
With this announcement Microsoft is bringing the power of the AKS engine to your local environment when deployed on the Microsoft Stack HCI platform, even in mostly disconnected scenarios. With promises of tight integration with on-premise Active Directory and other traditional data center technologies Microsoft is clearly driving a strategy that merges the old with the new.
The solution should allow organizations to keep their systems up to with the latest version of Kubernetes through the standard rolling updates etc.
The services however will not be free, but cost information has not been released.
Azure ARC Data Services
Azure ARC Data Services allow organizations to run Azure Data Services at the edge, in multicloud and on-premises environments using Kubernetes as the underlining supporting platform with or without an Azure connection.
The services supported with the launch of this public preview are Azure SQL Managed Instances and Azure Databases for PostgreSQL with the promises of more to come.
By leveraging Kubernetes on the backed, Microsoft can now deliver an always current version of your supported database engine and eliminate the need for database admins to ever worry about infrastructure upgrade, patching or clusters ever again.
AWS, GCP and all the standard cloud providers are fully supported allowing organizations to adopt the best of what Microsoft Data Services have to offer without relying on the Kubernetes clusters themselves to be running in Azure.
This service is currently in public preview and no additional pricing information has been released.
Azure Stack Edge Pro
Microsoft also announced the availability of Azure Stack Edge Pro appliances that not only includes a NVIDIA Tensor Core T4 GPU, but also includes support for virtual machines and Kubernetes deployments.
These devices allow organizations to bring standard compute to the edge, but further expands the Azure IoT footprint and edge computing models to virtually any location. With the support of Kubernetes, developers can rapidly deploy, update and optimize applications and tap into the machine learning capabilities that the new built in GPU(s) can unleash.
Although traditional Azure Stack Edge appliances did have the limited capabilities to support containers, the adoption of Kubernetes all the way on the edge once again shows that Microsoft is welcoming the a new “beyond-Windows” world with open arms.
Azure Stack Edge Appliances are procured via the Microsoft Azure Stack portal and billed on a per-month basis with a minimal once off shipping charge.
Microsoft wants Azure to be THE platform for your data
Norm Torvik, Practice Lead - Microsoft Data and AI
This year at Ignite, Microsoft had a lot of exciting news in the data and AI space to share: new services such as Azure SQL Edge and additions to Cognitive Services, as well as enhancements to existing services like Azure Databricks and Cosmos DB. The following are some of the highlights from this week.
This brings the SQL Engine to IoT devices to provide real-time analytics for IoT edge. Deployment options include online, offline and hybrid connectivity modes. SQL Edge provides support for X64 and ARM64 CPUs. Benefits include: always up to date (managed service), elastic scale, unified management and support for multiple clouds.
Azure Cognitive Services has introduced a series of new capabilities expanding about the existing Decision, Vision and Speech APIs.
The Decision API, Metrics Advisor Preview, is a new service that proactively monitors metrics and diagnoses issues. Built on Anomaly Detector (part of Azure Cognitive Services), Metrics Advisor monitors the performance of your organization’s growth engines — from sales revenue to manufacturing operations — through a powerful combination of monitoring in near-real-time, adapting models to your scenario, offering granular analysis with diagnostics and alerting.
The Vision API, Spatial Analysis (a feature of Computer Vision), helps organizations maximize the value of their physical spaces by understanding people’s movements and presence in near-real-time to create apps that can count people in a room, measure distances between customers, aggregate footfall in a store, understand dwell time in front of a retail display and determine wait times in queues.
The speech API now supports new functionality by adding support for containers to allow deployment across disparate environments.
Built on top of the latest Apache Spark codebase (v3.x), the new Delta Engine accelerates the performance through three components: an improved query optimizer, a caching layer that sits between the execution layer and the cloud object storage and a native vectorized execution engine that’s written in C++. These improvements make Azure Databricks 20x faster than Open Source Apache Spark.
Azure Cosmos DB now offers a serverless option for database operations. This new consumption-based model gives app developers a way to build and scale smaller apps and run tests without the commitment and cost of provisional throughput, making it ideal for small workloads with occasional traffic bursts and moderate performance requirements. It will be available across the MongoDB, Cassandra, Gremlin and Table set of APIs.
Feel free to comment on your favorite announcement below!