Dell Technology World is being held in Las Vegas, its typical home. Attendees will experience the latest and greatest Dell computing technology at the show. The one area of IT that conference goers will likely not be exposed to is advancements in networking. I attended Dell Technology World 23 and was hoping to see Dell make some headway in networking but outside of SONiC and some commodity switches, the portfolio significantly lags companies like HPE and Cisco.

For all the market-leading innovation Dell has had in PCs, storage systems and servers, it’s never managed to put together even a half-baked strategy in networking. Historically, this hasn’t impacted the company much as infrastructure was purchased and often ran in silos.

Artificial intelligence (AI) is changing that as it requires compute and networking to work together. This is why many vendors have invested in being a “full stack” provider. Nvidia acquired Mellanox, which has a robust line of AI switches. When HPE plunked down billions for Juniper networks, CEO Antonio Neri made it clear that the company was now network first. Cisco continues investing in its UCS server line to deliver end-to-end AI infrastructure. The lone holdout from the big IT vendors has been Dell.

Dell's lack of networking capabilities can be seen as a disadvantage in the AI era for several reasons:

The importance of data transfer

AI workloads are notoriously data-intensive, often involving massive datasets that need to be moved between different systems and storage solutions. Networks serve as the highways for these data transfers. Inadequate networking capabilities can turn data transfer into a major bottleneck, slowing down AI processes and reducing overall efficiency.

For Dell, traditionally a leader in hardware and server solutions, the absence of cutting-edge networking technology can be particularly detrimental. When an AI model requires fast access to large volumes of data, delays in data transfer can dramatically extend the time needed for training and inference, leading to slower innovation and development cycles.

Role in distributed computing

Distributed computing forms the backbone of many advanced AI applications, including deep learning and large-scale data analytics. These applications often require the computational power of multiple machines operating in concert. Networking is the glue that holds these systems together, enabling the synchronization and communication necessary for distributed operations. Dell's ability to offer solutions for distributed computing environments is compromised without strong networking capabilities. This can render their systems less attractive to businesses and researchers who rely on high-performance distributed computing for their AI applications.

Scalability challenges

In the AI era, scalability is not a luxury but a necessity. As AI models grow more complex, the need for infrastructure that can scale efficiently becomes paramount. Networking is a critical component of scalable AI systems, facilitating the interconnection of numerous servers and data centers. Dell's historical focus on hardware must encompass networking solutions to provide the scalability that modern AI workloads demand. Dell's infrastructure solutions may fail to meet the scalability requirements without networking, limiting their appeal to enterprises striving to expand their AI capabilities.

Real-time processing limitations

AI applications are increasingly being used for real-time decision-making and processing, from autonomous vehicles to real-time analytics and automated trading systems. These applications demand low latency and high-speed communication. Networking infrastructure that fails to meet these requirements can lead to delays and inefficiencies, undermining the effectiveness of real-time AI systems. For Dell, the lack of networking can be particularly disadvantageous, impairing their ability to offer support for real-time AI applications and consequently diminishing their market position.

Cloud integration shortcomings

The cloud has revolutionized the way AI workloads are managed, offering flexibility, scalability, and cost-efficiency. Networking is indispensable for seamless cloud integration, allowing for the smooth data transfer and computational tasks between on-premises systems and cloud platforms.

As many AI solutions increasingly rely on or integrate with cloud-based services, Dell's lack of networking strength can be a significant hindrance. It can inhibit hybrid cloud environments that many enterprises now demand, putting Dell at a disadvantage compared to competitors that offer more robust networking solutions integrated with their cloud offerings.

Final thoughts on Dell and AI

In the AI era, where data-handling speed, efficiency, and reliability are paramount, robust networking is indispensable. Dell's lack of networking capabilities stands as its Achilles heel, threatening to undermine its position in a market increasingly dominated by AI-driven innovation. Every aspect of AI infrastructure depends heavily on effective networking, from data transfer to distributed computing, scalability to cloud integration, and collaboration to real-time processing. Dell risks losing its foothold to more network-savvy competitors without significant improvements in this area, underscoring the urgent need to address this critical weakness.

Looking at infrastructure more broadly, Dell’s approach has been to sell networking as part of a solution with compute and storage rather than having a standalone network business. Dell’s roots are in computing, and it thinks compute first. However, the world is changing, and with the rise of mobility, cloud, IoT, remote work and AI, IT is now network-centric, hence Neri’s commentary on being network first. The world is changing, and Dell needs to change along with it either through internal development or via a merger or acquisition.