Many of us have seen the Will Smith movie “I, Robot” from 2004, getting a science fiction glimpse of what could happen if robots armed with artificial intelligence were in households in your neighborhood doing chores and making decisions between right and wrong. In the make-believe world of science fiction, we were given the opportunity to see AI as having limitless possibilities.
Fast forward to today and we see that not only does AI make an impact in the real world, but it is growing up in the cloud, and growing fast!
With the current work from home environment across the globe, cloud technology and datacenters are seeing tremendous use - and the outlook is very bright as the technology builds. According to International Data Corporation (IDC), the worldwide artificial intelligence software platforms market will approach $11.8 billion by 2023, with a 35.3% compound annual growth rate.
As the processing power required for AI and machine learning increases, it also surpasses the capability of traditional central processing units (CPUs). A chip called a graphics processing unit (GPU), initially used for computer gaming, is now at the forefront of the next wave of machine learning, says Forbes.
I sat down for a question-and-answer session with Maghen Hannigan, Director, Multi-Vendor Solutions for Tech Data, to learn more about how data centers are getting ready for AI with GPU’s.
Maghen Hannigan, Director,
What is your primary role at Tech Data?
I lead the Multi-Vendor Solutions team which includes a Program Management office and a Sales Execution team focused on our data center integration business. We work with both vendors and customers to bring together a vast array of technologies to deliver a fully integrated system to the channel. This includes secondary storage, hyper-converged servers, all the way up to application loaded converged systems.
Can you please explain artificial intelligence (AI) in layperson’s terms?
AI is the action of automation or to mimic expert thinking. Think decision tree or rule-based process. This concept has been around for years but has had major advancements with the evolution of machine and deep learning, which allow for more complex learning and decision making.
Basic AI mimics human behavior while machine learning uses algorithms to model data. Deep learning is AI with self-learning capabilities. It uses layers of algorithms to find similarities without being told what to do.
Can you tell us about some current business areas that AI is used in?
A few examples include:
Construction, specifically modeling and, architecture rendering; the special effects and video animation used in gaming and media production; and the landing simulations and the programs used to process images from space that astronomers and experts in the aeronautics industry use.
The finance industry uses this technology for a variety of things such as trading, fraud detection and protection, to determine loan qualifications, etc. There are also various AI internet services out there, including speech recognition programs.
Also, supercomputers with GPU technology are used for scientific research, including looking for a vaccine for the coronavirus. Some other common uses of this technology include manufacturing and navigation processes – and obviously robotics. The list goes on and on.
What are graphics processing units (GPUs) and why are they so important to the data center’s ability?
GPUs are computer chips that work in parallel with CPUs to process images. Traditional CPUs do not possess the computing power to process the data required for AI machine/deep learning. These are critical technologies needed to support digital transformation initiatives such as automated decision support, digital assistants or cognitive services.
What do Data Center managers/staff need to do to incorporate the GPU capability? Is it difficult and does it, require investment?
Many server and storage vendors have been incorporating GPU technology into their solutions to accelerate the support of AI workloads. Organizations / data center managers will have to assess the in-house skills and cost benefits of building this into their on-premises’ infrastructure or procure it through a cloud solution. The following questions need to be taken into consideration:
Do they have an AI strategy?
Do they have the skills in-house to test data and, ensure the data is clean?
Do they possess the capability to transform the data into insights that can be monetized?
Server/storage technology can be more expensive than traditional CPUs - but again, this is to enable new capabilities for an organization to deliver internal/external services.
Why is the message, “Get your data center ready for AI” important to share right now?
First of all, AI is making its way to the edge, the data center whether public cloud or on- premises will have to support these new edge devices and architectures.
Secondly, many digital transformation initiatives will rely on AI-enabled technologies to provide differentiated products and services to the market. Think digital assistants, autonomous cards, image modeling, etc.
How can Tech Data help vendor partners, chief information officers and data center managers with this endeavor?
We bring the customers and best-in-breed technology vendors together.
Through our sales enablement structure, we: help partners understand what technology solutions are available with use cases, promotions and positioning.
We also supply world-class technical support. Our Solutions architects help customers create solutions that will support the needs of their end users. We also provide solutions delivery support, data center procurement, integration and on-site installations with the ability for migration services.
Our AI Practice Builder leverages Tech Data's resident experts to walk partners, through the practice builder methodology step-by-step, demonstrating how to start a practice, providing pre-sales support and helping customers close opportunities. We understand the additional considerations around AI -such as the, - data integrity- and skillsets required and can help bring together what is needed.
As for cloud solutions, our Practice Builder and StreamOne platforms leverage AI products through the public cloud vendors. Customers and their end users can take advantage of these capabilities while simplifying the management and billing processes.
Speaking of cloud … behind every cloud is a data center, so what is the message to the cloud business applications community?
Cloud AI deployments will benefit with the increased demand of infrastructure as a service (IaaS) pay-as-you-go, metered offerings. Cloud providers are continuing to add to tool-kits for customers to leverage machine/deep learning capabilities quickly - services to build, train, deploy and enhance applications, to support workloads, etc. Advances in AI-enabled edge devices will accelerate the use of on-premises and public cloud AI-enabled solutions.
AI is already a part of our daily lives, whether we recognize it or not. Now with data centers implementing GPUs, deep machine learning is undoubtedly on the rise. As the gap between the science fiction world of “I, Robot” and real-world capabilities gets smaller, Tech Data is here to help.
IDC, Worldwide Artificial Intelligence Software Platforms Forecast, 2020-2024, Doc # US45724520, June 2020
About the author
Dwight Hawkins is a Marketing Communications Specialist for Tech Data in Tempe, Arizona. In his role, Dwight works on several external communications initiatives including the Authority blog, web content, partner newsletter, executive and presentation content and helping to build out thought leadership program and brand equity. He also hosts the video interview series Tech Data Fly-by. Prior to joining Tech Data Dwight served over 20 years in the United States Air Force garnering experience in public affairs, video and radio broadcasting, and strategic communication.