Cloud Hosting

Domino Data Lab is an AWS Partner Network (APN) partner that provides a central system of record for data science activity across an organization. The Domino solution delivers orchestration for all data science artifacts, including AWS infrastructure, data and services.

As part of the solution, Dominos platform leverages the scale, security, reliability, and cost-effectiveness of AWS cloud computing coupled with Amazon Elastic File System (Amazon EFS). Together they orchestrate all data science artifacts, such as AWS infrastructure, data, and services. This approach lets data science teams benefit from this flexible and collaborative research environment with automated workflows tracking model development dependencies for full reproducibility, including enterprise-grade governance, risk management, and granular cost controls.

In this post, we interview David Schulman, Director of Partner Marketing at Domino Data Lab, and explore the Domino Data lab Enterprise AI Platform solution to consider why centralizing data AI and machine learning operations (MLOps) initiatives into a single system of records across teams can help enterprises to work faster, deploy results sooner, scale rapidly and reduce regulatory and operational risk. In 2023, Domino surveyed artificial intelligence (AI) professionals with their REVelate State of Generative AI survey . The responders included AI professionals leading, developing, and operating AI across Fortune 500 companies. The survey reports that 49% plan to develop generative AI in-house, while 42% plan to fine-tune commercial models. Top limitations focused on security, reliability, cost, and IP protection. Consequently, 69% are worried about data leakage, with both top leadership (82%) and IT (81%) being especially concerned.

According to Ventana Research, Through 2026, nearly all multinational organizations will invest in local data processing and infrastructure and services to mitigate against the risks associated with data transfer. Hybrid cloud environments complicate the operationalizing of AI/ML at scale by creating silos across data, infrastructure, and tooling. Challenges include data science teams that are prevented from collaboration because of siloed data, processes, and tools. Non-standardized, non-repeatable, ad-hoc bespoke workflows result in sprawl on individuals computers across systems. Data and compute resources are distributed across cloud and on-premises data centers, causing unconnected environments and silos. There are also hidden costs resulting from data scientists spending time on DevOps and infrastructure management tasks to prevent underutilized infrastructure due to idle, always-on, and over-provisioned resources.

Over years of partnership, Domino and AWS have worked together to assist organizations, such as Johnson & Johnson (JnJ), in reducing analysis time by 35% for data scientists[1]. This involves integrating Dominos data science platform with essential AWS services, such as Amazon EFS. Amazon EFS provides analytics storage with shared file access to data scientists. Applications include open-source genomics and Shiny, and Domino Data Lab, run on an Amazon Elastic Kubernetes Service (Amazon EKS) cluster. Amazon EFS provides access to a fully managed petabyte-scale file system supporting genomics sequence data at 500 TB. More recently, JnJ has further scaled data science across their hybrid and multicloud environment, adopting AI infrastructure strategies straddling on-premises data centers and the cloud to address concerns over cost, security, and regulatory compliance. Lilly also centralizes data science to drive value across the healthcare value chain, as discussed last spring on a panel at NVIDIAs GTC AI Developer conference.

Dominos Enterprise AI Platform integrated with key AWS services provides a unified, collaborative, governed, and end-to-end MLOps platform. The solution orchestrates the complete ML lifecycle, providing easy access to data, preferred tools, and infrastructure in any environment. By sharing knowledge, automating workflows across teams, and tracking all changes and dependencies, Domino guarantees complete reproducibility while fostering collaboration. It also helps maintain peak model performance in production while making sure of enterprise-grade model governance and cost-savings. Domino can be deployed into VPCs, or it is available as a SaaS offering on AWS Marketplace. Attributes include:

Domino is Kubernetes-native and can be deployed on Amazon EKS for ease of management across hybrid environments. This enables cloud-native scalability, multi-cloud portability, reduced costs through elastic workloads paired with underlying hardware resources, and simplified administration by integration with existing DevOps stacks.

Domino can run on a Kubernetes cluster provided byAWS Elastic Kubernetes Service. When running on EKS, the Domino architecture uses AWS resources to fulfill theDomino cluster requirements as in Figure 1.

Figure 1: View Domino Documentation for Domino on Amazon EKS

Seamless collaboration and knowledge sharing is a requirement for data science teams. First, Domino Datasets, integrated with Amazon EFS provide high-performance, versioned, and structured filesystem storage in Domino so that data scientists can build curated pipelines of data in one project and share them with colleagues for collaboration. Amazon EFS enables the sharing of data pools among multiple instances that were previously isolated from one another. This increases data science team productivity because Domino not only tracks snapshots of data used to build models, but all of the underlying code, packages, environments, and all supporting artifacts for full reproducibility providing rich file difference information between revisions. Additionally, customers such as JnJ value the Amazon EFS storage class feature which enables them to automatically move data from Amazon EFS Standard to Amazon EFS Infrequent Access. By automating the process of moving data to long-term, cost-effective storage, the customer successfully reduced their storage cost.

With Amazon EFS, storage capacity is elastic, growing and shrinking automatically as you add and remove files to dynamically provide storage capacity to your applications as needed. With elastic capacity, provisioning is automatic, and youre billed only for what you use. Amazon EFS is designed to be highly scalable both in storage capacity and throughput performance, growing to petabyte scale and allowing massively parallel access from compute instances to your data. This makes it the perfect data science platform foundation for organizations such as JnJ to reduce analysis time by 35%.

Why does this matter? With Amazon EFS, data science teams are empowered to:

Second, Domino Data Sources act as a structured mechanism to create and manage connection properties for external sources such as Amazon S3, Amazon Redshift, and a variety of sources. This reduces DevOps work for data scientists, as they get desktop-like data store connectivity without needing any coding.

Flexible model deployment options support diverse business and operational requirements. Models developed in Domino can be exported for deployment in Amazon SageMaker for scalable and low latency hosting, while models from SageMaker and SageMaker Autopilot can be accessed and monitored inside Domino for drift and prediction performance. Models can be deployed to the cloud, in-database (deploy to Snowflake or Databricks for predictive analytics), or to the edge (Domino supports NVIDIA Fleet Command). Models can be deployed for both batch and real-time predictions at scale, while Domino Model Sentry controls model validation, review, and approval processes for an additional governance layer.

Hybrid cloud support is a necessity for enterprise data science teams, and Domino Nexus acts a single-pane of glass for all data science workloads across hybrid and multi-cloud environments.

Figure 2: Domino/AWS architecture

A Domino Nexus deployment consists of a control plane, a Kubernetes cluster hosting Dominos core platform services (deployed on Amazon EKS as in the preceding Figure 2) and one or multiple data planes: distinct Kubernetes clusters that run a small set of Domino services that can execute workloads. These can be deployed in any cloud region, across multiple clouds, or in on-premises data centers.

Figure 3: Nexus Hybrid Architecture | AWS Cloud Control Plane (US East Region), AWS Cloud Data Plane (EU Central Region)

As shown in Figure 3, Users connect to the Domino control plane through a browser connection, while users connect directly to the data plane where they are doing their model development work in a Domino Workspace. Amazon Elastic Load Balancer (Amazon ELB) allows ingress to Domino control plane services from data planes.

This architecture (Figure 4), eliminates the possibility of inadvertently transferring region-locked data. It also allows data scientists to seamlessly burst to the cloud if they run out of on-premises compute capacity.

Figure 4: Source: Nexus hybrid Architecture Source

Dominos Enterprise AI Platform is proven to deliver an average of 722% ROI in three years (average NPV and ROI based on study of Domino customers, using Domino Business Value Assessment). This is achieved by 2x more models delivered with the same resources, in the same amount of time. A 40% reduction in data scientist time wasted waiting for IT support, doing DevOps, or duplicating work. A 40% reduction in IT and cloud infrastructure costs over three years. Reduced risks of revenue loss from violations of regulations or reputation issues.

Want to go deeper into these metrics? Learn more about Domino and cost-effective AI. Read the following case studies from enterprises with Dominos platform deployed on AWS:

Although generative AI gets all of the attention, large language models (LLMs) are in fact just models. And although there are many intricacies in operating generative AI at scale, such as prompt engineering, model fine-tuning, and inference/hosting (well save that all for another post), the following best practices of scalable, enterprise AI remain the same:

Flexibility is required by modern enterprises, who need to build, deploy, and operate AI at scale across a variety of complex architectures. In addition, storage plays an important role, and Amazon EFS provides a cost-effective, elastic, and highly performant solution for your ML inferencing workloads. You only pay for each inference that you run and the storage consumed by your model on the file system. Amazon EFS provides petabyte-scale elastic storage so your architecture can automatically scale up and down based on the needs of your workload without requiring manual provisioning or intervention.

To learn more about how Dominoon AWS can accelerate responsible AI initiatives? Downloadthe Strategies and Practices for Responsible AI TDWI playbook for insight on a proactive approach on identifying and mitigating business, legal, and ethical risks to create trust and deliver tangible business value.

Visit Domino on AWS Marketplace.

[1] AWS Storage blog Johnson & Johnson reduces analysis time by 35% with their data science platform using Amazon EFS

Read more:

Optimizing enterprise MLOps in the cloud with Domino Data Lab and Amazon Elastic File System | Amazon Web Services - AWS Blog

In collaboration with NICHD, NIH, and external stakeholders, ODSS is building a federated, secure research and specimen data ecosystem that will measurably and rapidly facilitate data and specimen sharing by NICHD-funded researchers and increase access to shared data and specimens for the entire research community.

The NICHD Data Ecosystem includes people, data, processes, and technologies that align with the NICHD Strategic Plan and that support NICHD communities data science and sharing needs. In addition, ODSS is collecting user stories to describe these needs, within the following contexts:

The NICHD Ecosystem Use Case Library on GitHub tracks user stories, use cases, efforts, and associated documentation to improve capabilities across the ecosystem.

ODSS is currently assessing all data repositories that NICHD researchers use to share data or access shared data for secondary use. This assessment and the communities user stories will inform NICHDs approach to improving sustainability and interoperability across the ecosystem to best support the institutes needs. The office will also use information from the assessment to update the NICHD Data Repository Finder .

More:

Office of Data Science and Sharing (ODSS) | NICHD - Eunice Kennedy Shriver National Institute of Child Health and ... - National Institute of Child...

Tom Siebel in 2017 in Redwood City, California. Courtesy University of Illinois

Tom Siebel, the billionaire head of artificial intelligence (A.I.) company C3.ai, is funneling $50 million into data science research and education at his alma mater. The tech CEOs donation to the University of Illinois Urbana-Champaign will transform its computer science department into the Siebel School of Computing and Data Science.

The new school, pending approval from the University of Illinois Board of Trustees and Illinois Board of Higher Education, will be part of the universitys Grainger College of Engineering. Its establishment exemplifies the University of Illinoiss dedication to pushing the boundaries of knowledge and fostering collaborative solutions to global challenges, said Graingers dean Rashid Bashir in a statement. This transformative gift will empower our faculty and students to lead the next generation of technological advancements, further solidifying our position as a world-renowned institution.

The donation will help the university keep up with the rapid pace of research into computing and data science, which have become a key aspect of engineering alongside physical sciences and mathematical sciences, Bashir told Government Technology magazine. Siebels funds will reportedly support research project opportunities, in addition to funding a new building for the school.

Siebel grew up in Chicago and earned three degreesa bachelors in history, an MBA and a masters in computer sciencefrom the University of Illinois Urbana-Champaign before going on to work as an executive at tech company Oracle. He subsequently created the software firm Siebel Systems and sold it to Oracle in 2006 for $5.8 billion.

He currently has an estimated net worth of $3.7 billion and has since 2009 run C3.ai, which specializes in utilizing enterprise A.I. to aid large-scale companies. It counts the U.S. Air Force, Department of Defense, Shell and Con Edison (ED) among its customers.

We are thrilled to partner with the University of Illinois to establish the Siebel School of Computing and Data Science, said Siebel in a statement. By supporting cutting-edge research and fostering innovation, we hope to empower future generations of leaders in technology and society, driving positive change in our world.

This isnt the first time Siebel has used his fortune to give back to the educational institution. He donated $32 million in 1999 to create a computer science center and another $25 million in 2016 to create a design center, in addition to giving $100 million to its science and engineering programs in 2007.

In 2019, Siebels C3.ai launched a generous education benefit that reimburses eligible employees who pursue masters degrees in computer science with a focus on data science at the University of Illinois Urbana-Champaign. In addition to covering the total degree cost, the company offers employee graduates a 15 percent salary increase, $25,000 cash bonus and stock awards.

See the original post here:

Billionaire Tom Siebel Donates $50M to the University of Illinois for Data Science Research - Observer

Photo by ANGELA BENITO on Unsplash

Theres a core concept in machine learning that I often tell laypeople about to help clarify the philosophy behind what I do. That concept is the idea that the world changes around every machine learning model, often because of the model, so the world the model is trying to emulate and predict is always in the past, never the present or the future. The model is, in some ways, predicting the future thats how we often think of it but in many other ways, the model is actually attempting to bring us back to the past.

I like to talk about this because the philosophy around machine learning helps give us real perspective as machine learning practitioners as well as the users and subjects of machine learning. Regular readers will know I often say that machine learning is us meaning, we produce the data, do the training, and consume and apply the output of models. Models are trying to follow our instructions, using raw materials we have provided to them, and we have immense, nearly complete control over how that happens and what the consequences will be.

Another aspect of this concept that I find useful is the reminder that models are not isolated in the digital world, but in fact are heavily intertwined with the analog, physical world. After all, if your model isnt affecting the world around us, that sparks the question of why your model exists in the first place. If we really get down to it, the digital world is only separate from the physical world in a limited, artificial sense, that of how we as users/developers interact with it.

This last point is what I want to talk about today how does the physical world shape and inform machine learning, and how does ML/AI in turn affect the physical world? In my last article, I promised that I would talk about how the limitations of resources in the physical world intersect with machine learning and AI, and thats where were going.

This is probably obvious if you think about it for a moment. Theres a joke that goes around about how we can defeat the sentient robot overlords by just turning them off, or unplugging the computers. But jokes aside, this has a real kernel of truth. Those of us who work in machine learning and AI, and computing generally, have complete dependence for our industrys existence on natural resources, such as mined metals, electricity, and others. This has some commonalities with a piece I wrote last year about how human labor is required for machine learning to exist, but today were going to go a different direction and talk about two key areas that we ought to appreciate more as vital to our work mining/manufacturing and energy, mainly in the form of electricity.

If you go out looking for it, there is an abundance of research and journalism about both of these areas, not only in direct relation to AI, but relating to earlier technological booms such as cryptocurrency, which shares a great deal with AI in terms of its resource usage. Im going to give a general discussion of each area, with citations for further reading so that you can explore the details and get to the source of the scholarship. It is hard, however, to find research that takes into account the last 18 months boom in AI, so I expect that some of this research is underestimating the impact of the new technologies in the generative AI space.

What goes in to making a GPU chip? We know these chips are instrumental in the development of modern machine learning models, and Nvidia, the largest producer of these chips today, has ridden the crypto boom and AI craze to a place among the most valuable companies in existence. Their stock price went from the $130 a share at the start of 2021 to $877.35 a share in April 2024 as I write this sentence, giving them a reported market capitalization of over $2 trillion. In Q3 of 2023, they sold over 500,000 chips, for over $10 billion. Estimates put their total 2023 sales of H100s at 1.5 million, and 2024 is easily expected to beat that figure.

GPU chips involve a number of different specialty raw materials that are somewhat rare and hard to acquire, including tungsten, palladium, cobalt, and tantalum. Other elements might be easier to acquire but have significant health and safety risks, such as mercury and lead. Mining these elements and compounds has significant environmental impacts, including emissions and environmental damage to the areas where mining takes place. Even the best mining operations change the ecosystem in severe ways. This is in addition to the risk of what are called Conflict Minerals, or minerals that are mined in situations of human exploitation, child labor, or slavery. (Credit where it is due: Nvidia has been very vocal about avoiding use of such minerals, calling out the Democratic Republic of Congo in particular.)

In addition, after the raw materials are mined, all of these materials have to be processed extremely carefully to produce the tiny, highly powerful chips that run complex computations. Workers have to take on significant health risks when working with heavy metals like lead and mercury, as we know from industrial history over the last 150+ years. Nvidias chips are made largely in factories in Taiwan run by a company called Taiwan Semiconductor Manufacturing Company, or TSMC. Because Nvidia doesnt actually own or run factories, Nvidia is able to bypass criticism about manufacturing conditions or emissions, and data is difficult to come by. The power required to do this manufacturing is also not on Nvidias books. As an aside: TSMC has reached the maximum of their capacity and is working on increasing it. In parallel, NVIDIA is planning to begin working with Intel on manufacturing capacity in the coming year.

After a chip is produced, it can have a lifespan of usefulness that can be significant 35 years if maintained well however, Nvidia is constantly producing new, more powerful, more efficient chips (2 million a year is a lot!) so a chips lifespan may be limited by obsolescence as well as wear and tear. When a chip is no longer useful, it goes into the pipeline of what is called e-waste. Theoretically, many of the rare metals in a chip ought to have some recycling value, but as you might expect, chip recycling is a very specialized and challenging technological task, and only about 20% of all e-waste gets recycled, including much less complex things like phones and other hardware. The recycling process also requires workers to disassemble equipment, again coming into contact with the heavy metals and other elements that are involved in manufacturing to begin with.

If a chip is not recycled, on the other hand, it is likely dumped in a landfill or incinerated, leaching those heavy metals into the environment via water, air, or both. This happens in developing countries, and often directly affects areas where people reside.

Most research on the carbon footprint of machine learning, and its general environmental impact, has been in relation to power consumption, however. So lets take a look in that direction.

Once we have the hardware necessary to do the work, the elephant in the room with AI is definitely electricity consumption. Training large language models consumes extraordinary amounts of electricity, but serving and deploying LLMs and other advanced machine learning models is also an electricity sinkhole.

In the case of training, one research paper suggests that training GPT-3, with 175 billion parameters, runs around 1,300 megawatt hours (MWh) or 1,300,000 KWh of electricity. Contrast this with GPT-4, which uses 1.76 trillion parameters, and where the estimated power consumption of training was between 51,772,500 and 62,318,750 KWh of electricity. For context, an average American home uses just over 10,000 KWh per year. On the conservative end, then, training GPT-4 once could power almost 5,000 American homes for a year. (This is not considering all the power consumed by preliminary analyses or tests that almost certainly were required to prepare the data and get ready to train.)

Given that the power usage between GPT-3 and GPT-4 training went up approximately 40x, we have to be concerned about the future electrical consumption involved in next versions of these models, as well as the consumption for training models that generate video, image, or audio content.

Past the training process, which only needs to happen once in the life of a model, theres the rapidly growing electricity consumption of inference tasks, namely the cost of every time you ask Chat-GPT a question or try to generate a funny image with an AI tool. This power is absorbed by data centers where the models are running so that they can serve results around the globe. The International Energy Agency predicted that data centers alone would consume 1,000 terawatts in 2026, roughly the power usage of Japan.

Major players in the AI industry are clearly aware of the fact that this kind of growth in electricity consumption is unsustainable. Estimates are that data centers consume between .5% and 2% of all global electricity usage, and potentially could be 25% of US electricity usage by 2030.

Electrical infrastructure in the United States is not in good condition we are trying to add more renewable power to our grid, of course, but were deservedly not known as a country that manages our public infrastructure well. Texas residents in particular know the fragility of our electrical systems, but across the US climate change in the form of increased extreme weather conditions causes power outages at a growing rate.

Whether investments in electricity infrastructure have a chance of meeting the skyrocketing demand wrought by AI tools is still to be seen, and since government action is necessary to get there, its reasonable to be pessimistic.

In the meantime, even if we do manage to produce electricity at the necessary rates, until renewable and emission-free sources of electricity are scalable, were adding meaningfully to the carbon emissions output of the globe by using these AI tools. At a rough estimate of 0.86 pounds of carbon emissions per KWh of power, training GPT-4 output over 20,000 metric tons of carbon into the atmosphere. (In contrast, the average American emits 13 metric tons per year.)

As you might expect, Im not out here arguing that we should quit doing machine learning because the work consumes natural resources. I think that workers who make our lives possible deserve significant workplace safety precautions and compensation commensurate with the risk, and I think renewable sources of electricity should be a huge priority as we face down preventable, human caused climate change.

But I talk about all this because knowing how much our work depends upon the physical world, natural resources, and the earth should make us humbler and make us appreciate what we have. When you conduct training or inference, or use Chat-GPT or Dall-E, you are not the endpoint of the process. Your actions have downstream consequences, and its important to recognize that and make informed decisions accordingly. You might be renting seconds or hours of use of someone elses GPU, but that still uses power, and causes wear on that GPU that will eventually need to be disposed of. Part of being ethical world citizens is thinking about your choices and considering your effect on other people.

In addition, if you are interested in finding out more about the carbon footprint of your own modeling efforts, theres a tool for that: https://www.green-algorithms.org/

The rest is here:

Environmental Implications of the AI Boom | by Stephanie Kirmer | May, 2024 - Towards Data Science