In the last two years, more organisations in India launched their digital transformation journeys and today, India is ahead of most countries in the adoption of digital technologies. A Dell Technologies study titled ‘Digital Transformation Index’ confirms that. But organisations continue to grapple with the challenge of managing unstructured data and large data sets.
In an interview with DIGITAL CREED, Charles Sevior, CTO, APJ, Unstructured Data Solutions, Dell Technologies explains how data management solutions and high-performance storage solutions can counter the data challenge.
Charles Sevior, CTO, APJ, Unstructured Data Solutions, Dell Technologies
In the last two years, more organisations in India launched their digital transformation journeys and today, India is ahead of most countries in the adoption of digital technologies. A Dell Technologies study titled ‘Digital Transformation Index’ confirms that. But organisations continue to grapple with the challenge of managing unstructured data and large data sets.
In an interview with DIGITAL CREED, Charles Sevior, CTO, APJ, Unstructured Data Solutions, Dell Technologies explains how data management solutions and high-performance storage solutions can counter the data challenge.
He also talks about the four pillars of Digital Transformation and how Dell Technologies is using the pillars to help organisations with Digital Transformation.
GPU’s are increasingly being used for high-performance computing (HPC) and Dell has partnered with NVIDIA to offer HPC solutions. Sevior offers some applications and explains how the business world is gravitating towards HPC for applications such as the connected car, in the automobile industry.
Excerpts from the interview:
DC: How are the disruptions happening around us? What are the technologies that you have for enterprises to deal with challenges?
CS: There is a lot of disruption already happening around us and to study that, one of the things that we did was conduct research called the Digital Transformation Index. It’s a detailed study of 4,000+ global CIOs across enterprise companies. We did this study in 2016, and then we repeated it in 2018. We do it in a 2 – 3 years’ gap, and it’s an assessment of where your organization is being represented in the survey, and where it is sitting across the five categories – Digital Laggards, Digital Followers, Digital Evaluators, Digital Adopters, Digital Leaders. Are you a digital leader or a digital native? It’s everything you are doing designed around data and integration of data in your tools. Or are you a laggard, the traditional manufacturing industry or a traditional enterprise that’s not doing anything special. Or are you somewhere between?
Dell Technologies has invested in the digital transformation of organizations, in helping organizations to become more efficient, derive more value out of the data and become a data-driven organization. Therefore, this is a measure where we see the Fortune 500 companies and the 2000 behind them, where do they sit in the spectrum of maturity in terms of understating digital. A lot of the leaders often led by young CEOs, understand from the Digital Transformation Index, which is educating CEOs and CFOs in business about the value of data and its importance of being a data company.
The findings of the report suggest that the percentage of Digital Leaders and Digital Adopters has increased in India. Whereas globally, there’s been no progress at the top, and almost 4 in 10 (39%) businesses are still spread across the two least digitally mature groups on the benchmark (Digital Laggards and Digital Followers). Hence, definitely, there is a shift and we expect this curve will get pushed in the right direction as time goes on. Another interesting finding is that India was the most digitally mature country in the world. In fact, 91% of the Indian business leaders believed that digital transformation should be more widespread, compared to 78% globally.
IMAGE CREDIT: Dell Technologies
DC: So, in the context of digital era, what is it that Dell Technologies provides in terms of this transformation?
CS: At Dell Technologies, we have a process that helps our customers achieve success and we call these the four pillars of transformation, which are: Digital Transformation, IT Transformation, Workforce Transformation and Security Transformation.
The first pillar is Digital Transformation itself. There is a dire need of new-age applications amongst the organisations, which can help them in their journey to become a successful digitally transformed business. Second, this Digital Transformation must be supported by IT Transformation. One cannot imagine a digitally transformed world and applications without a robust technology infrastructure supporting it. The third pillar is Workforce Transformation. The infrastructure should be able to support the workforce of the future. It should support their needs and aspirations, and at the same time, the new way of working – whether it is mobile or remote working. The fourth and most critical pillar is Security Transformation. It is all about the overall digitization of the business and ensuring security of all aspects of an organisation. We live in a more connected world, with more devices being connected every day, and that means more security risks are opened up to hackers. Therefore, security is of the utmost importance.
So, as a technology conglomerate, we are well-positioned to provide solutions across all of these and for that, we have a portfolio of companies like Secureworks, Pivotal and RSA.
Moreover, we have been introducing the concept of ‘data first’ approach as it is the most valuable thing for any company. Data is actually recognized in the valuation of the company through its goodwill and investments and the information that helps the company in their trading, their business partners and their direct relationship with consumers. Therefore, it becomes important to carefully manage the data and not just manage and protect, but also focus on its back-up.
For unstructured data, the problem is the huge amount of data growth and sprawl. A lot of the companies have investments in specific projects and they might get investment for some initiatives and data may get stored in some new platform in different data silos, in some other line of businesses.
DC: What is the major challenge that companies face while managing these large data sets?
CS: For unstructured data, the problem is the huge amount of data growth and sprawl. A lot of the companies have investments in specific projects and they might get investment for some initiatives and data may get stored in some new platform in different data silos, in some other line of businesses. Therefore, the company might not be aware of all the information that is sitting in different parts of the business, posing a big problem. Moreover, the challenge we face while working with our customers is how they can see the data crossed over different lines of businesses in different siloes, how do they manage the security in the governance, and what sort of platform do they need to successfully mine the data.
DC: What are the problems around data management?
CS: Although data management is a great idea, what an organization needs is a terrific start at it. They should get some high-level infrastructure to store data, some high-level analytics system and some data scientists/trained people which will make the analytics easy. However, as per our observations, we saw that nearly 50% of a resource’s time is just basic data management and preparation. So what we try for our customers is to make this process a little less time consuming by closing the gap between laborious data processing, and sowing a traditional data warehousing environment.
DC: Are you talking about putting all the data in a Data Lake or residing it in one place?
CS: There is a shift away from storage management to data management that is making everything as one data pool. Around 80% or more data is unstructured and this is photos, texts, scanned documents, audio in call centres, videos that are coming from self-driving cars or kiosks, and all these sorts of things. This 80% of data copy is read by machines since it’s too much for humans to sit down and do. Most companies are often storing the data on tape backup like a record and that becomes frozen or locked away. One can use it if it’s in tape backup after restarting the disk.
We need to classify the value of data, but any data that is operational and has context that can be used in your training and machine learning, becomes the baseline for machine learning.
DC: How has Storage life cycle management evolved?
CS: Storage life cycle management was known as hierarchical storage management (HSM). This has changed over the years since no data is unimportant. Data should be classified as important or completely not necessary to keep. If it is not necessary to keep, delete it. It’s always advisable to delete it because we don’t know what its use is. We need to classify the value of data, but any data that is operational and has context that can be used in your training and machine learning becomes the baseline for machine learning. Therefore, such applications should be kept online which is why we have a platform called Dell EMC ECS (object storage) as a part of our unstructured data solutions portfolio.
There is a classification in tiers and we have different types of storage. We have high performance, low latency, fast storage for data which needs fast storage. We have the Isilon range that covers the high-performance inference data, the training data that is being used to build an artificial intelligence model. Hence, a lot of the training data can be put on the ECS public storage which has much lower cost, and which is designed for less frequent access.
We are seeing a shift in transformation in the automotive industry. There are huge changes being made and we’ve already seen the impact of Uber and Ola as compared to traditional taxies and owned vehicles.
DC: So, at the top, you have the All flash storage, for high performance?
CS: In Isilon, we have tiers, either all flash or disk-based storage, and disk is placed at a different performance level. We also have ECS, which is also disk-based but it is designed like an object. Hence, Isilon is for fast online high-performance storage and ECS is for online storage.
For an instance, in a case of healthcare, looking at an image taken from your body to see if it is cancer, what type of cancer is it and then classifying it into a particular sort of cancer, etc. will be quicker in an automated way. Therefore, across different industries/sectors, we can see different use cases where machine learning and deep learning techniques that are developed help to do these things much quicker in an automated way.
Another example where we are seeing a shifting in transformation is of the automotive industry. There are huge changes being made and we’ve already seen the impact of Uber and Ola as compared to traditional taxies and owned vehicles.
DC: Is Dell going to be using GPUs in its systems for high-performance computing?
CS: We have a really close partnership with NVIDIA. Many of our customers in automotive, for example, who are developing these high-performance platforms or healthcare or in aerospace, have been investing in these high performing GPU systems. We’ve been selling GPU cards in our Dell workstations and services for many years, and we have the ability now with Isilon for this really high-performance storage that goes well with the deep learning algorithm and multi-threaded algorithm. So, it’s very powerful. We can simply have customers using the PowerEdge servers to connect using parallel IO, connect the Isilon storage, that’s where we’re training it with historical data, and they can learn very fast at high performance.
At Dell Technologies World, we announced a new Dell Technologies server that implements up to 10 Nvidia Tesla V100 GPUs inside the server and it is just another step up of the density and power of the Dell servers. We have also partnered with NVIDIA to sell Dell servers with NVIDIA cards in them or we can partner with the NVIDIA servers which many of our customers are buying.
DC: What would be the typical application for this?
CS: The typical application before was for training data for the self-driving cars, all that data which is coming in is basically used to train the AI system as to what to do when it sees this certain object and recognize it, etc. A lot of this is around image recognition but can also be around audio recognition, processing speech, different accents, different dialects, etc. We have a ready solution for deep learning and this is a combination of Dell servers within video cards and Isilon storage and networking. Additionally, we have a system that is a reference architecture of Isilon, which uses video servers together with Dell storage and networking to provide the customer with what they need.
Recently, at Dell Technologies World, we announced a new Dell Technologies server that implements up to 10 NVIDIA Tesla V100 GPUs inside the server and it is just another step up of the density and power of the Dell servers. We have also partnered with NVIDIA to sell Dell servers with NVIDIA cards in them or we can partner with the NVIDIA servers which many of our customers are buying, either option.
