Dell Technologies World was all about AI; what about security?

As a cybersecurity industry analyst, I spend most of my time working with security companies, but at Dell Technologies World last week, I was immersed in conversations about AI.

Though Dell isn't a cybersecurity company, it offers several high-quality managed security services focused on helping its customers keep their operating environment safe and secure. Services include security program assessment and maturity development, zero-trust strategy development, managed detection and response, vulnerability management and traditional incident response services.

While rich in security talent to support these offerings, Dell's engineering investment in these areas primarily focuses on integrations, implementation and unified management. The company works closely with various security providers, including Zscaler, Okta, CrowdStrike, Halcyon, Palo Alto Networks, Deep Instinct and Index Engines -- all of which were at the conference.  

I spent plenty of time with Dell's services team and the many security vendors present at the event, but I was more focused this year on how and where Dell is delivering AI in support of the massive transformation that companies are going through as they AI-enable their operations.

Dell is all-in on AI, which dominated the conversation throughout the week. The Dell AI Factory was at the center of it all, and Dell laid out a strategy for every company to plot a course for their own AI journeys. While running fast, Dell was quick to admit that it is learning along with its customers but made a pledge to be transparent and to communicate what the company learns with all by sharing reference playbooks that can be utilized within the AI factory.

Dell's AI strategy is anchored on five tenants:

1. Data is the differentiator.
2. Data gravity matters -- building models where your data already lives is more efficient, effective, and secure.
3. No one-size-fits-all.
4. Things are moving fast, so using an open, modular architecture is the only practical approach to leverage all the knowledge and available innovation.
5. A broad and open ecosystem is required to keep up.

Dell believes most enterprises will ultimately end up using their own small language models that combine their own data, augmentation and publicly available data. These models will be optimized over time, fueled by intelligent data pipelines to ensure the models are primed with fresh data.

But Dell says the traditional data center is simply not equipped to support all this. A different scale of computational compute power is necessary, together with high-speed I/O for files and object data stores; data storage and management subsystems to prepare, clean and validate data used to build models; and a high-throughput, low-latency network backplane to enable a massively distributed compute environment.

AI workloads, referred to as "elephant workloads," require massive processing, power, cooling and network capabilities. The first three weren't a surprise to me, but I hadn't thought much yet about the network requirements needed to support the massive parallelization needed to scale up a cluster to build large language models. This distributed computing architecture can only work if the backend network can enable these systems to work together. Dell is partnering with Broadcom to tackle this problem.

Much focus was on the AI data pipeline that fuels the AI rocket ship. Dell made it clear that keeping the data close to the source is the best approach, versus pushing massive amounts of data into the cloud for processing. In a customer session I attended, this proved to be true, as they found that moving the data to the cloud for processing created delays that resulted in longer build times for models. They reported that moving the processing on premises where the data resides delivered up to five times improvement in the speed of model creation. This of course is powered by Dell PowerEdge XE servers.

As Dell pivots its agenda to an all-in AI strategy, a new opportunity to repatriate data and on-premises processing has the potential to dramatically change the growth trajectory for Dell, leaving me to wonder if this will cause IT leaders to reconsider cloud strategies, or will this be a net add in spending, further growing overall IT spend in support of the AI revolution.

Dell has also added professional services offerings to support customers' AI journeys, including strategy development, use case definitions, help with building the models, including preparing and cleansing the data, education to build out internal knowledge to enable self-sufficiency, and infrastructure planning and design.