By Roland Barnes
Most, if not all, of our customers face challenges around their data management, with today’s very real hybrid multi-cloud landscape affecting enterprises of all sizes. Sometimes seen as a bit of a long-standing cliché, back in 2017, The Economist published a story titled, "The world's most valuable resource is no longer oil, but data”. It is apparent this cliché has become so true to all our customers, as they strive to gain true business value from their data and become more competitive in their industries.
As customers look to unlock this value from their data, many are faced with the familiar ice-berg effect in that they only know and trust the top 20% of their data for analysis. Many customers have never implemented a robust Information Lifecycle Management (ILM) policy to categorise and understand all their rafts of data, which leads to the 80% below the water often being totally unknown, not trusted, and unnecessary, with multiple copies of the same data existing when not all required. Indeed, more than half of the data (53%) at the average organization is stale and taking up space without adding value. Logicalis believes and recommends our customers all try and implement this ILM discipline, often summarised by our data management life cycle, as shown below!
Without reliable ILM practices in place, our customers frequently have this iceberg landscape of data across multiple environments such as:
- On premise silos of storage
- Multiple vendors and infrastructure
- SAN, File and HCI environments
- Data spanning multiple Public Cloud subscriptions
- Remote branch/office locations
- Edge data
One study by IBM found that businesses spend 70% of their time finding data, with only 30% of time allocated to the analysis, and we have seen this too. Once we speak to customers it soon becomes abundantly clear of the challenges they face bringing all these disparate factors together so they can focus their time on the data analysis that drives business growth.
Customers also need to ensure that the staff employed to analyse their data are best utilised for their expertise instead of constantly trying to ensure the data is trusted pre- and post-processing. Expensive FTE resources such as Data Scientists and Business Analysts shouldn’t spend more time doing non-productive tasks such as cleansing data rather than the detailed analysis and training of results.
Modernise your storage
Modernise is a commonly used term in our ever-changing world of hybrid multi-cloud and is often directed at the application layer for legacy workloads. However, to get better value from their data, customers should also aim to modernise their data estate too.
Most enterprises use multiple tools to analyse their data for business value and competitiveness, with three of the main areas being:
- Big Data & Analytics (BDA)
- Artificial Intelligence (AI)
- Machine Learning (ML)
Applications supporting the above areas all rely on the data that they are processing being trusted, performant, governed, accessible and cost effectively stored, not in a too dissimilar theme to the ILM approach mentioned earlier. If we were to class the above three application categories into a simple AI and data naming convention, our message to share with our customers would be that there is no AI without an Information Architecture (IA) – yes more anagrams!
No AI without IA
So, what do we mean by this? Back in 2017, IDC predicted that by 2020 there would be a global spend of $10 billion on storage just to support AI projects. With the pandemic having rapidly accelerated many businesses shift to cloud automation; we see this as a massive growth area. In this digital world we know customer’s data is growing at frightening rates, but it must be stored somewhere.
Right now, the sprawl of data across multiple platforms prevents users from being able to easily access and control it. To mitigate this challenge, the Information Architecture (IA) is key to ensuring our customer’s data challenges are met, overcome and can finally deliver true business benefits and outcomes, but also contribute to better ILM discipline for their data. This concept of the Information Architecture can be thought of as the “ladder” for AI and data.
One of Logicalis’ strategic partners is IBM - hold that thought please and dispel any myths about legacy and expensive now – I promise I can prove those myths wrong! IBM’s view on this ladder can be summarised by the following simple graphic.
Collect is the first rung of the ladder, it’s the foundation of everything as data is the fuel that powers AI, ML & BDA. Data can become trapped or stored in a way that makes it difficult or cost prohibitive to maintain or expand, collect or process, often resulting in duplications; in one study conducted by M-Files, 83% of employees reported that they have recreated existing documents, as a result of being unable to find the original versions in their company network. Customers need the ability to ensure that the data is in the right place, at the right time, accessible to the relevant people and applications, but most importantly at the right cost profile too – true policy driven management of data.
IBM Storage for Data and AI makes data simple and accessible for a hybrid multi-cloud infrastructure, but more importantly also as Software Defined solutions offering customers choice of HW and Cloud infrastructure.
One key component worthy of awareness and understanding is Spectrum Scale, a swiss army knife with many use cases in the world of AI/ML and BDA, under-pinning many of the world’s largest HPC clusters. However, it can start as small as is required and scale to yottabytes.
Organise is our next step on the AI Ladder. As we have highlighted AI and analytics can only be as good as the data it relies on. Indeed, issues with keeping data current can cost businesses in the United States over $600 billion per year. Hence, businesses must fully understand what data they have so they can leverage it for AI/ML and BDA and other organisational needs, including compliance, data optimisation, data classification and data governance.
Industry experts have been arguing for years that for optimal business growth and protection of privacy, data management needs to become more dynamic, responding to business needs through an extensive library of data while also protecting systems from breaches. IBM Storage for Data and AI offers solutions to allow data updates in real-time, which means as new data is ingested, it is automatically updated to the storage catalogue and can leverage the policy engine for optimised placement and meta data tagging. Imagine knowing more about your data spanning those silos on and off premise and across multiple cloud environments!
Spectrum Discover is one of the solutions that can help customers overcome these data organisational challenges purely as a Software Defined solution, offering choice of platform on or off-premises, but scanning and understanding the whole iceberg!
Analyse is us now stepping further up the AI Ladder. We’re now ensuring that the data we are using for AI/ML and BDA is much more trusted, governed and relevant to our goals of getting value from the data. The Analyse rung is the foundation of being able to understand your data and glean the insights and information you need in order to build and scale AI/ML models and algorithms with trust and transparency. High performance, massive scalability, multi-protocol access, Cloud and Tape tiering capability, and automatic policy driven placement of data are key for the analysis of the curated and trusted data.
Again, IBM Storage for Data and AI delivers solutions to allow these capabilities with Spectrum Scale, offering multi-protocol access, un-paralleled scalability and performance vs the competition along with IBM Cloud Object Storage (ICOS) allowing massive capacities for low cost highly encrypted and always available Object/S3 data storage.
Infuse becomes available once we have reached the top rung of the ladder, the benefits of getting there can be expanded across the organisation. For organisations to truly operationalise AI/ML and BDA, they need to be able to put the results to work in multiple departments and within various processes. From payroll, to customer care, to marketing, drawing on predictions, automation, and optimisation, to help advance their business agenda.
Business challenges can become an opportunity to explore, understand, predict by bringing an AI/ML and BDA infrastructure to customers and their entire organisation. Logicalis are happy to share how IBM Storage for Data and AI can help get there.
Thank you for reaching this far. There’s a lot of information above, but in essence customers need to modernize their data storage by facilitating change as a process that starts with reducing cost and complexity and ends with faster business results and faster time to market.
Here’s one last picture I’d like to share 😊
Logicalis are passionate about solving our customers’ business challenges and can share our extensive experience and roadmap for successful data transformation to help you achieve successful outcomes to your data challenges.
If you would like to learn more about Data Management, you can download our free eBook here. If you would like to understand more by speaking to one our of specialists, please email firstname.lastname@example.org and we will set up an initial call.