2023 - the year in Infrastructure

George Lawton Profile picture for user George Lawton December 27, 2023
Summary:
The year's highlights in IT infrastructure.

infra

2023 was a coming out year for digital twin technology that is helping firms of all shapes and sizes digitalize representations of physical assets and weave them into new workflows. 

All of the major product lifecycle management vendors have migrated their solutions to the cloud to streamline integration across data silos. AI innovations are also helping to pave better digital threads that can further weave in more types of data across departments, subsidiaries, and partners. New 3D standards have also emerged to streamline digital twin workflows further. 

Other innovations in test intelligence, hallucination mitigation, and time calibration are starting to demonstrate practical value for enterprises. 

In 2024, enterprises pressed to address sustainability issues will likely turn to these advancements to meet their goals faster and more cost-effectively. 

How the Industrial Metaverse will be different to the consumer one - and why it matters

I predict Industrial Metaverse platforms will gradually improve interoperability in response to industry demands. It will be easier to share data and processes between different categories of vendors like mapping and Industrial design. However it may take a while for better sharing within any particular category of tools. Look at how tough multi-cloud is today, even after twenty years. Cloud providers tend to charge a lot more to move data out than in. The Industrial Metaverse may not be much different.

Why: We are still in the early days of building the Industrial Metaverse, but vendors like Siemens, Dassault, PTC, and NVIDIA are already delivering real enterprise value. The consumer metaverse is getting the most hype, but the enterprise version is already starting to play a seminal role in meeting sustainability goals, building physical products more quickly, and connecting the dots across data and process siloes. 

Software AG lays out progress bridging data app silos

It is amazing how fast this technology is evolving. It is amazing how we are using LLMs [Large Language Model] and augmenting it with our own data. We just assumed so much has been done, so why not leverage that? Before, with a Google search, you just get a number of links. What the generative AI is doing well is giving answers to questions like, ‘How can I map these fields from system A to system B?’

Why: In the short run, enterprises are likely to see more value from applying to LLMs to get their data house in order for all kinds of AI, machine learning, and analytics use cases rather than just rushing into building a better chatbot. That’s Software AG Chief Product Officer Dr. Stephan Sigg talking about the value of LLMs in transforming and integrating data. He was surprised at the early success in applying generative AI to data integration challenges and expects to see it automate 70-80% of the most common integrations in the near future. 

Grieves and Vickers - digital twin challenges and opportunities

As it turned out, the digital twin predicted the point of failure to within three percent and for a few tens of thousands of dollars rather than hundreds of millions for the physical prototype. However, the existing culture mandated spending hundreds of millions of dollars on the physical prototype.

Why: Digital twins started out as a way to improve product lifecycle management across siloes. Although the term sometimes gets conflated with modeling or simulation, the real value lies in figuring out disparate processes using the right combination of ontologies, integration tools, data management, and physics simulation.

AOUSD - how USD is becoming the formal standard for the Industrial Metaverse

I think it's a really big deal. Again, not just for the film industry, where it was kind of born. But for industries well beyond film. Whether it's immersive 3D content, interactive experiences, new spatial computing platforms, or scientific and industrial applications, OpenUSD will become the fundamental building block on which all 3D content will be created.

Why: Today, there are dozens of different 3D data formats, which can lead to many integration challenges in managing workflows that span physical products, machine vision, autonomous systems, IoT data, and new kinds of physical automation. Industry efforts to harmonize data formats across these use cases will help enterprises automate and speed up these workflows internally and with partners and customers. 

Altair – the marriage of digital twins and LLMs

We anticipate the first wave of the integration of LLMs with digital twins to be through providing a natural language interface to navigate and explore digital twins. The next phase will be using LLMs to generate data related to the content of digital twins that can be used in the modeling and control of digital twins.

Why: Digital twins and LLMs are two very complementary technologies. The structured models in digital twins can reduce hallucinations in LLM, while the conversational interfaces of LLMs can help navigate the complexities of digital twins. Today, new coding tools are using LLMs to speed development using testing tools to root out hallucinations at the source. Digital twins will extend these testing tools to other use cases, not just physical ones. In Altair’s case, this pairing is showing value in banking and financial analysis of complex portfolios. 

Why electric grid bottlenecks present new opportunities and costs

Taylor suggests that more practical grid decarbonization approaches may need to focus on the grid in its entirety rather than just the power lines. For example, to look at how investments, regulations, and technologies around smart grids, energy storage, and multi-vector alternatives could complement shortfalls in the power grid. A multi-vector approach might mean transforming some electricity into hydrogen and then distributing it over existing gas pipelines, where it could be converted into electricity or heat closer to the point of use.

Why: In the rush to NetZero goals, it's tempting to focus on solar, wind, heat pumps, and electric cars while glossing over the wires that connect these things. That’s an oversight since power grid upgrades are expected to cost about $6 trillion over a decade. And not just for the wires. Digital technologies will be essential in streamlining the provisioning process, building stakeholder agreement, and managing the new grid. This is a tremendous opportunity for digital innovation. 

How Harness is using AI to safely streamline operations

Everyone wants to ship very fast, but the number one reason they cannot is because you don't know if you will break something. And that process can take a long time. AI can really help in learning if an update is going to break something.

Why: Harness demonstrates the value of getting your data house in order as a foundation for building better AI that can solve practical infrastructure problems. In this case, it is for creating the infrastructure for test intelligence that reduces the time developers spend waiting from two hours to ten minutes. Harness only started adding LLMs once they had a solid data infrastructure in place to ensure the results were right and valid. 

The many types of AI hallucinations (Clue - it's not that simple!)

Enterprises and governments have struggled with data quality and accuracy since before the dawn of computers. But concerns about AI hallucinations have picked up steam with the rapid adoption of Large Language Models (LLMs), like ChatGPT, that can make up believable bullshit on an impressive scale. Not only that, but they can convey these with a computer-generated straight face.

Why: The Cambridge Dictionary recently named hallucinate the word of the year. This is not an entirely new problem in enterprise IT. The UK Post Office scandal sparked by a hallucinating IT system in 1999 is still being sorted out in the courts. But the rush to deploy LLMs is starting to amplify the problem. Now, vendors and researchers are developing various metrics, tools, and workflows to help identify and mitigate these issues as enterprises begin to roll out LLMs. 

New services are automating world-scale digital twins - here's how

But just building a 3D version of the world isn’t enough. You have to be able to pull data out of it and understand it. You want to know how many cars are where and how cities are growing. Once you can understand the world, you can start making better decisions… You can start answering really hard problems at a global scale about supply chains and insurance.

Why: It took Google over fifteen years to create a Street View map that covers 98% of the places people live. Innovations in satellite imagery, 3D processing, and semantic ontologies are allowing a new wave of innovators to replicate this feat daily. Moreover, these modern variants don’t just show what it looks like - they can help assess insurance claims, predict commodity flows, streamline supply chains, calculate solar potential, and plan infrastructure projects.

Clockwork – discovering wasted bandwidth between the nanoseconds

If these [quartz] clocks were temperature controlled, you can get down to the parts per billion. So, it'll be some small number of nanoseconds per second. Now, those kinds of clocks and network interface cards could easily be in the few hundreds of dollars to possibly up to $1,000 on their own. And the next level is rubidium clocks, which are three to five grand, and then cesium. As you add these costs to the raw cost of a server, you're piling up the costs across a large data center. So, it'd be nice if we could do it without having to resort to that. And that's more or less what we do.

Why: Who would have thought that better clock synchronization software could cut networking and cloud infrastructure costs by ten percent or more? It also reduces errors in transaction processing systems and databases by orders of magnitude. Down the road, it could allow teams to take advantage of new decentralized architectures that were not previously practical. 

Loading
A grey colored placeholder image