Microsoft Build is a major event for Avanade, connecting with product and service engineers, and allowing us to share with our clients the latest insights to platforms and services we’ve been working on with Microsoft.

In the current climate, Microsoft boldly chose not to cancel their flagship developer event, opening virtually to the whole world. Avanade hosted a round-the-clock live chat on Microsoft Teams with our people around the world, from Australia, to the Netherlands.

From the very beginning, we saw the scale of Microsoft’s vision when Satya touted Azure as the world’s computer.

Microsoft has launched their big book of news – but at 84 pages, it’s a long read. We’ve watched the announcements and combed through the revelations, from Azure Analytics, to Zero-friction applications: here’s our (much shorter) take.

Fluid Framework – distributed collaboration, and an open source commitment
Chris Lloyd-Jones, Emerging Technology Engineering Lead
Organizations have workforces distributed across the globe, using collaboration tools to share knowledge, complete their work, and get things done.

Productivity tools were designed for a face-to-face world. They’ve evolved, but not revolved, trapping information in documents, limiting information traceability. These tools can’t handle scalability, and real time updates, promoting process over people.

Microsoft is open sourcing the Fluid Framework to power collaborative distributed applications- when you start a paragraph in Word, you can access that content with others in any other tool, on any device, and even in an application you want to develop.

I expect to see an explosion of tools, unlocking knowledge from individual documents, powering experiences designed for a distributed world.  

Cristy Stone, Emerging Technology Research Lead
This is an interesting move on Microsoft’s part (despite joking references to Apple’s early “OpenDocs” measure), not least because the open sourcing of the components will likely have a net-positive effect on adoption even outside of the Microsoft ecosystem. And these types of advances in technology tend to lead to new ways of thinking and more creative output, something companies are hungry for in a challenging economy and amongst ever-more automation.

Why care? Organizations can benefit from increased uptime and highly responsive platforms, and employees are empowered to be more productive, and engaged.

ONNX Runtime: Improve machine learning performance, and deploy across multiple platforms
Fergus Kidd, Emerging Technology Engineer
Training machine learning models is costly, and time consuming. Creating human parity language understanding models consumes huge amounts of resources.
Advances in AI require hugely parameterized models to be more generally useful than what is already available.

The ONNX Runtime is a framework that accelerates ML-training and inferencing on any platform, such as TensorFlow or PyTorch. Microsoft announced that the key optimization techniques used for the Turing model will be available from June.

Microsoft demonstrated extremely powerful outputs from these large models, such as high-level language understanding, document searching, contextual awareness, and document summarization.

Why care? These optimizations allow ML engineers to create better models, with fewer resources.

Open AI – Azure Supercomputer
Bruno Capuano, Regional Innovation Lead for Canada, and Microsoft AI MVP
Microsoft has a new supercomputer that’s the fifth fastest in the world- to do this, they teamed up with OpenAI, a startup cofounded by Elon Musk.  The focus of this new computer is to reach the next level of AI. Usually machine learning experts develop models focused on a single task, this new supercomputer will help the creation of multitasking models.
 
And, if you like technology, the numbers behind this computer are amazing: the OpenAI supercomputer is a single system with more than 285,000 CPU cores, 10,000 GPUs, and 400 gigabits per second of network connectivity for each GPU server.

Cristy Stone, Emerging Technology Research Lead
OpenAI’s ambitions are no less lofty than true artificial general intelligence, and they’ve made headlines with Rubik’s-cube-solving robotic hands and strategy game winning bots. This is the first time Microsoft has shared details on the outcome of the $1 billion investment they made in the organization last year, and it looks like they’ve stayed focused on their strategy of enhancing productivity for their customers (after first doing it for themselves). This new access to sophisticated models and powerful compute could be especially compelling for companies with an established AI pipeline and infrastructure as they’re under pressure to deliver more in challenging economic times.

Why care? More compute capability will lead to more powerful AI implementations across the board at a lower cost.

"Possible intrusion in the west wing..." - What if your building could talk to you?
Dan Benitah, London Digital Studio Technical Lead
The indoor mapping service opens a whole new world of possibilities. I’m especially excited about the related announcement of new changes to Azure Digital Twins, for integration with the new indoor maps service.

The combination of these new tools will allow rapid deployment of fully-fledged applications with realistic maps of our customers facilities, allowing at a glance visualizations of desk occupancy, open door alerts, and overall utilization of their resources.

These announcements bring smart buildings into easy reach for our customers, enabling rich relevant building scenarios, and are perfect for supporting organizations as they figure out their COVID-19 transition plans.

Cristy Stone, Emerging Technology Research Lead
This is an important step to breaking down the boundaries between the physical and the digital; the initial focus will be on building management for owners and tenants (which will reap big benefits post-COVID), but we anticipate that smart companies will also being integrating the virtual version of the building with virtual versions of the employees. 

Why care? It’s now easier to digitize your internal spaces, and the barrier to entry for smart buildings has been reduced.

Azure Synapse Link
Kyle M. Hale, Data and AI Solution Lead, Southwest Region
Transactional engines like rows; analytical engines like columns. In that small sentence lies the enormous, decades-long struggle of DBAs, EDW architects, ETL wranglers, data modelers, and report developers to perform that transformation, with the ultimate goal of minimizing time to insight. Go ask your favorite data guy about the sleepless nights spent pondering, “We already know what the data looks like where it is, and what we want it to look like where it will be. Isn’t there an easy button I can press?” and you’ll quickly learn that this is a large, complex problem. There’s even a term for this Nirvana state – Hybrid transactional/analytical processing, or HTAP. With all of the innovations in data management and storage solutions coming on the market in recent years, HTAP is finally moving off the hype curve and into reality.
 
Enter Azure Synapse Link, Microsoft’s new take on the HTAP “easy button” (it’s literally a button, if you believe the Build announcement!) By enabling Azure Synapse Link, you get automatic syncing from your transactional systems (Azure CosmosDB is the sole supported source today, with others to follow) into Azure Synapse Analytics. They’ve done the heavily lifting behind the scenes to manage the alchemy of rows to columns for all of the disparate dataset workloads and patterns they’re expecting to see, and then passing on this shared resource to all of us.
 
Here’s to hoping this leads to a little less sleepless nights for us tired data architects, and more focusing on the insights our data is generating!

Why care? Azure Synapse has the potential to reduce complexity increase scalability of database management.

 

Read part 2 of our Microsoft Build 2020 event summary.