In the era of digital intelligence, what is the greatest differentiated competitiveness of enterprises and organizations?
The answer is unequivocally: data. With the rapid development of generative AI technology, the development of new quality productivity has become an industry consensus, and more and more enterprises realize that data is the foundation of all operations, and it is the key for enterprises to embrace the AI wave, achieve qualitative changes in productivity, and build differentiated competitiveness.
In fact, the advent of generative AI has not only made the role and value of data more prominent, but also brought a series of new data storage requirements and challenges. Therefore, enterprises urgently need to build a strong digital foundation through smart data infrastructure to provide a steady stream of "fuel" for AI training and inference, so as to comprehensively drive AI applications to the ground in the transformation and upgrading of digital intelligence.
Among many manufacturers, Lenovo Lingtuo relies on the comprehensive advantages of "advanced technology and local research and development", and through a series of software, hardware and solution innovations, it helps Chinese enterprises and organizations build intelligent data infrastructure and fully embrace the AI wave in an all-round way. As Lin Yousheng, senior director of product management and marketing at Lenovo Lingtuo, said: "The core of new quality productivity is artificial intelligence. The next three years will be an important window period for the integration of AI and business, which will bring earth-shaking changes. AI needs to talk to data, and data will be the decisive point of AI applications. ”
Lin Yousheng, Senior Director of Product Management and Marketing, Lenovo Lingtuo
AI readiness requires hard work
The computing power is valuable, and the "price" of storage power is higher.
As industries begin to fully embrace artificial intelligence and combine generative AI technologies such as large models with business scenarios, more and more enterprises realize that generative AI is essentially an engineering problem for large-scale, high-quality data and efficient data processing, and data quality determines the performance, generalization ability, and application effect of large model algorithms. The acquisition of high-quality data is closely related to data infrastructure, which needs to meet the stringent requirements of data storage, training, analysis, and inference of generative AI in terms of performance, capacity, scalability, and stability.
For example, large models are rapidly moving towards multimodality, and multimodal data varies greatly in terms of scale and data form, requiring sufficient capacity and data processing capacity. For another example, with the continuous improvement of the parameter size of large models, the scale of AI clusters is also rising, and 10000 or even 10,000 calorie clusters mean that high-frequency and large-scale CheckPoints will be generated during the training process, which will exponentially increase the reliability challenge of data storage. Another example is that with the convergence of large models and business scenarios, a large number of inference requirements are generated, followed by large-scale parallel and complex IOs.
Therefore, in the AI era, data infrastructure first needs to have "hard work" to fill the performance level of data processing, provide stable, reliable, and high-performance advanced storage capacity for AI applications, and support the performance requirements of the whole process of AI data processing.
There is no doubt that Lenovo Lingtuo is setting a new benchmark for data infrastructure in the AI era. Take Lenovo Lingtuo's new NetApp AFF A all-flash series as an example, including the AFF A70, A90 and A1K all-flash unified storage systems, which can be scaled up to 64 controllers, providing 40 million IOPS, up to 1 TB/s data throughput, and proven 99.9999% data availability.
Lin Yousheng bluntly said that high performance is a necessary capability for advanced storage power, and it is also one of the core elements of smart data infrastructure. The NetApp AFF A all-flash series features a new bus architecture design and chip-level hardware acceleration to deliver superior performance for any data type and demanding workload.
In fact, with the application of generative AI, multi-cloud, big data analytics and other technologies, enterprises are facing a far more complex data environment than in the past. If the "hard work" such as performance raises the lower limit of data infrastructure, then the "soft power" of data management is expected to continuously raise the upper limit of data infrastructure.
Simplicity is inseparable from soft power
The complexity of the data environment is a huge challenge that many enterprises have to face today.
Taking generative AI as an example, the data types and processing requirements in the stages of data acquisition, data preprocessing, model training, and application deployment are different, and there are still a large number of data collaboration and management work.
Similarly, the rapid development of artificial intelligence has also promoted the rapid development of cloud collaboration scenarios, and the flow and collaboration of data in the cloud, edge, and end are becoming normalized.
In addition, ransomware has led to frequent cyber attacks, and the rise of generative AI such as large models has made large model data a new target for attack. As we all know, model training in vertical industries relies heavily on their own industry data, and once an attacker steals model data, it will undoubtedly cause huge losses to the enterprise.
Therefore, a unified storage platform that can simplify data management, support the full lifecycle data management in the AI era, and effectively protect data will become a necessary condition for the future development of enterprises.
As the pioneer of unified data storage, NetApp, one of Lenovo's parent companies, has more than 30 years of experience in the field of data storage, and its ONTAP software platform has been an industry leader in data management, data collaboration and multi-cloud integration. With the continuous upgrade and iteration of the ONTAP 9.15 software platform, Lenovo Lingtuo is expected to inject more powerful "soft power" into the data infrastructure of Chinese enterprises.
First of all, ONTAP, as a proven unified storage platform, naturally supports compatible NAS, object storage, block storage, big data and other protocols, and a single platform can carry the data storage and management needs of the entire AI life cycle, helping enterprises completely eliminate data silos and greatly driving the implementation of AI applications.
Based on the groundbreaking data fabric concept, ONTAP truly integrates data flow, collaboration, and management, simplifying complex data environments and management tasks, and truly realizing data-driven services.
Third, ONTAP has strong data protection capabilities, and data protection functions such as SnapMirror and FlexCache have been tested in various harsh business scenarios for a long time. In addition, ONTAP embedded machine learning models enhance autonomous ransomware detection to prevent cyber threats, achieving more than 99% accuracy in detecting ransomware attacks, further improving data infrastructure data protection.
"Simplifying data management and trusted data security is also a core element of smart data infrastructure." Lin Yousheng introduced.
Build an ecosystem in the AI era
McKinsey Research predicts that generative AI is expected to contribute $7 trillion to the global economy by 2030. Among them, China, as a highland for AI R&D and application, will share one-third of the total benefits of generative AI through strategic investment.
There is no doubt that the Chinese market's emphasis on artificial intelligence and new quality productivity will greatly promote the digital and intelligent transformation and upgrading of vertical industries. At the same time, the accelerated integration of technologies such as generative AI, big data analytics, and multi-cloud into business scenarios has also made the importance of the ecosystem more prominent to a certain extent.
Lenovo attaches great importance to building a future-oriented ecosystem around new quality productivity. For example, generative AI is accelerating into vertical industries, but generative AI is technically complex, difficult to deploy and develop, and still faces a very high threshold for popularization in the industry. According to Gartner's Top Storage Recommendations to Support Generative AI, three-quarters of organizations with generative AI training data will deploy a single storage platform by 2028, and the "GenAI in a box" converged storage solution will help lower the barrier to entry for generative AI in vertical industries.
In fact, Lenovo Lingtuo's two parent companies have long realized that ecological cooperation is conducive to promoting the implementation and popularization of generative AI. To that end, the two parent companies are introducing the NetApp AIPod with Lenovo solution, a new converged infrastructure solution dedicated to generative AI retrieval augmentation and inference use cases, featuring Lenovo's high-performance ThinkSystem servers with NVIDIA GPUs, NVIDIA Spectrum-X networking, and NetApp AFF storage.
"AIPod is more like a reference architecture, which does not limit the flexibility of the user, and the user can choose the right storage according to the actual application." Lin Yousheng introduced.
In addition, Lenovo Lingtuo is also working vertical industry solutions, focusing on the needs of customers in medical, manufacturing, transportation, finance and other industries, and working with partners to create out-of-the-box industry solutions.
"Lenovo Lingtuo's biggest advantage lies in local R&D + advanced technology, with a complete local R&D and service support system, and industry-leading scientific and technological capabilities. Lenovo Lingtuo hopes to assume the responsibility of an ecological operator, combine and interconnect computing power, storage power, and data around the smart data infrastructure, and work with partners to continuously match industry applications and help enterprises develop new quality productivity in the era of digital intelligence. Lin Yousheng finally said.