Artificial intelligence (AI) has been a hot topic since Stanley Kubrick introduced us to the HAL 9000 in “2001: A Space Odyssey.” From self-driving car news in real life to movies about robots that mirror human consciousness, people have long been fascinated by the idea of a machine that can actually think. It makes sense, then, that machine learning and analytics have garnered the same level of fascination since they have become a large part of AI as we know it today. Meanwhile, data storage is an equally vital component of today’s machine learning technology, but it’s rarely the star of the latest sci-fi blockbuster.
Storage isn’t the sexiest thing in the world, but today’s analytics-based AI wouldn’t exist without it. As enterprises increasingly prioritize machine data analytics and management, they need to ensure IT infrastructure can support such initiatives in order to produce ROI. Below are three things about data storage every IT pro interested in AI should know:
1. AI and machine learning involve a ton of data.
AI projects are undeniably “cool,” in an obvious sense and a technical manner – but to accomplish the goals they set, they need to correlate and mathematically analyze massive data sets. This data tends to generate on the edge of the network – where people and machines all live, and the data must be frequently accessed with high availability and low latency for analysis. Often, maintaining this level of access comes at a high cost – which is why the companies making waves in the AI space are often major enterprises and incumbent vendors, rather than resource-strapped startups.
2. In the cloud-focused future, there are still some tough problems to solve.
Developers, not traditional IT professionals, are driving the creation and adoption of next-generation architectures that aim to leverage public clouds while supporting AI, machine learning, machine data analytics and more. Trying to spin a “Goldilocks” architecture from scratch that can provide the performance, latency, and cost characteristics that these workloads require is far from a trivial effort. When problems such as these arise, IT has an opportunity to produce innovative solutions rather than throw up roadblocks. Getting involved, and driving discussion about alternatives, is the best way to remain relevant.
3. “Storage is dead” is one of the industry’s greatest myths.
Compared to conferences in the last decade, VMWorld 2016 had a distinctly different atmosphere. Vendors that once dominated enterprise IT recently made major changes within their organizations, and those actions sent waves throughout the technology market. Sometimes, resulting effects from these changes help populate one of the industry’s biggest myths – that storage is dead. On the contrary, data storage has just embarked on some of its biggest changes yet.
For enterprises and startups alike, machine learning and AI will only continue to take root in companies’ interests, priorities and eventually, roadmaps. Data storage will need to support that journey every step of the way – although the systems of the past were never built to handle the massive amounts of data generated by AI projects, Internet of Things (IoT) information and machine data analytics. To stay alive, storage is evolving to meet the needs of such initiatives.
Read five things every machine data analytics user should know about data storage