Data Protection is More Than an Insurance Policy – It’s an AI Must-Have
- April 5, 2018
In a world reshaped by digital transformation, data has become an integral piece of the decision-making process within an organization. All business models today are built around data, enabling leaders to make big decisions to increase revenue, decrease cost, and reduce risk. However, too often organizations view the protection of that data as an overhead or an insurance policy, rarely ever looking at the data they’ve actually stored. That data has a lot of value in it that if accessed…could be a game changer when it comes to business insights and intelligence.
Organizations need to start thinking about data management by considering the following questions:
And of course, there are industry trends that will shape the way you think data management and protection.
Over 70 percent of enterprise companies are expected to leverage AI by the year 2021. Innovation in specialized hardware is accelerating deep learning capabilities in the data center, translating into business insights that drive growth through historical data that can now be derived from machine learning algorithms.
Building AI models and instantiating them is a vital investment – which of course, comes with risk. That data is incredibly valuable. The reality is that at some point your valuable data may be lost, damaged, corrupted, or compromised – the equivalent of taking 500 of the smartest people at your company and having them disappear.
Data protection, and the accessibility of that data, has become significantly more important (than ever) for the future of the data center – and everywhere it lives – including the cloud and at the edge. This is more than just an insurance policy – this is critical in ensuring your most valuable data assets are there when you need them to drive growth and a competitive market position.
Data protection today is a very complex task. There are a lot of manual configurations to consider, a need for specialized administrators, and an intense process for accessing historical data.
Fast forward into the future: we would like to have a fully autonomous system that isn’t just storing data arbitrarily, but using compute capacity to gain insights and develop a model to create the brain for your business. This would instantiate data through which the algorithms, results, trends, and trained models could extract the most value. Think about what this could do!
When you consider data protection – protecting and trusting where your data lands – it’s not just about making sure it is resilient. It’s about making sure that all that effort you expended to develop an AI learning system over an extended period of time – the code and the data – lands somewhere that it is actually protected. The system must be intelligent enough to know on which location to store the data, what meta-data you need to produce the data, and be able to predict possible data loss events and prepare for them in advance. And intelligence drives simplicity.
Most protection data is kept on spindles. As new media storage is becoming cheaper, high capacity QLC flash devices continue to expand in capacity without increasing the price. Arriving to the market, these systems are able to store secondary data (i.e. the backup data) on a much faster storage device and still gain the price and efficiency of a data protection system.
The future protection of storage will be integrated into an AI system and thus insights of past data will be available.
Being able to gain insights directly from the secondary protection system will also allow reducing some of the load and capacity on the primary storage arrays, allowing a wider adoption of new storage technologies like storage class memory (SCM) for primary storage.
As data becomes critical to the organization, a simplified, coherent data protection plan is required for new application development mechanisms too. Not only that, but more and more enterprises are moving to use hybrid cloud and multi-cloud strategies, meaning data not only resides on premise or on a single cloud, but rather on multiple different clouds. Over 20 percent of the enterprises believe that they will use more than five separate clouds in the future. This makes management protection of the data an even more complex task. But it also opens opportunities. In addition, the IoT devices that generate huge amounts of data pose a challenge for proper data management and protection strategies. I plan to explore all of this and more in future posts.
As we can see, the value of business data paired with the reality of artificial intelligence has created a new world order. Our research shows that data protection initiatives were one of the most common initiatives undertaken by companies that are looking to transform and modernize their IT. The outcome? A robust, current environment adapted to keep up with the likes of the data generated by artificially intelligent infrastructure. Data protection is not merely an insurance policy. It’s a must-have to make big decisions and insights in order to stay competitive in this digital landscape.