By Paul Speciale, CMO, Scality
The last few weeks of 2021 are dwindling away and here at Scality, we’ve been using this time to reflect on what we’ve seen in data storage in the past year and what’s to come.
Here’s what won’t change: The proliferation of data will continue to explode, and almost every organization across every sector needs to have a defined strategy for grappling with all this data.
So what does 2022 have in store? Join us as we gaze into our crystal ball.
Increasing cyber attacks demand sophisticated cyber protections to be built in
It’s not hyperbole to say that ransomware has reached epic proportions in the past couple of years. Cyber criminals are wise to the value that corporate data holds and they are doing everything they can to get at it. That means commercial storage solutions will need to be designed with more sophisticated, integrated mechanisms for earlier detection, prevention and — ultimately — recovery from attacks that delete, modify or encrypt stored data.
Storage solutions will be combined with advanced application-level, server and network security mechanisms to provide corporations with end-to-end solutions against cyber attacks across their IT stacks. Threats are continuing to evolve; stopping ransomware attacks from happening isn’t realistic, so IT managers will have to deploy solutions that can help detect and recover from these attacks earlier and more efficiently. For data storage solutions, early detection abilities can come in the form of new AI/ML based techniques that analyze anomalous access patterns — or are focused on certain data. Flagging these conditions can give administrators an earlier warning, before significant damage is done.
Recovery solutions in data storage can entail classical snapshot technologies and more granular versioning capabilities. New object storage systems have taken data immutability to even higher levels by implementing object locking, along with data retention policies. These effectively render data impervious to deletion or modification for the specified period. The ultimate solution may reside in smart ways to “air gap” sensitive data – either to a remote location or to a securely connected system. We anticipate such solutions will continue to rise in sophistication and be offered by more and more vendors.
AI/MLOPS will become a standard part of enterprise and midrange storage products
In addition to massive data growth, the data storage industry continues to grapple with a skills shortage. It’s been estimated that storage administrators will have to manage 50 times more data in the next decade, with only a 1.5X increase in the number of skilled personnel.
Automation and intelligence can meet these challenges and help IT managers and operations address a range of tasks – from basic systems and alert monitoring, to predictive analytics and capacity planning, to sophisticated data workflow management.
We expect to see more integration of AI/MLOps into large-scale data storage offerings to address the skills shortage and help administrators offload and automate processes. It can also help with finding and reducing waste as well as increasing overall storage management efficiency.
MLOps can monitor and provide predictive analytics on common manual tasks (i.e. capacity utilization, pending component failures and storage efficiencies.) Higher value tasks can include determining application IO and performance patterns to automatically adjust Quality of Service (QoS) and the underlying resources to deliver the needed service levels. These innovations wouldn’t be possible without the application of ML techniques and their ability to consume and “train” from extremely granular system logs and event data during real-time operations.
Digital sovereignty concerns will spur an industry of regional service providers
In Europe, there is a lot of discussion about the dependence on U.S. and Asia-based semiconductor manufacturers, infrastructure vendors and the big three cloud vendors – AWS, Microsoft Azure and Google. Growing concern over this dependence has driven the introduction of new policies aimed at enhancing the continent’s (and the individual nations’) strategic autonomy in the digital field. But digital sovereignty isn’t solely a European issue. In the U.S., increasing dependence on China for certain technologies, especially silicon chip manufacturing, among others, has drawn similar concerns.
Data is a highly valued enterprise asset in today’s digital economy and we predict the creation of new localized services to address the growing concerns around the independence of sovereign data. Companies are struggling to keep track of the location of their data amidst the growth of the biggest worldwide cloud players – not to mention the status of their compliance with local data sovereignty regulations they’re subject to. Outsourcing and delegating IT services to global cloud service providers for the sake of economics, agility and flexibility does not absolve them of their compliance obligations. Turning a blind eye to their data sovereignty issues is just not an option. Therefore, this will usher in an industry of local/regional service providers offering sovereign cloud services to captive markets by ensuring the data stays within specified borders.
On-premises data centers will remain an enterprise priority for the foreseeable future
Decentralization of IT services, applications and data has been an ongoing trend over the last decade. Applications and data have long been moving from corporate data centers to public clouds. However, the enterprise data center is far from dead. We predict that corporations will maintain their investments in corporate on-premises data center infrastructure even as cloud adoption continues for reasons of control, performance and cost-efficiency. And we expect this to continue for the foreseeable future.
This will lead to a new level of sophisticated IT management capabilities to optimize multi-data center, multi-cloud application and data management solutions. Data storage and management is a multi-site hybrid-IT problem.
Most enterprises today have realized that a smart, balanced approach to applications and infrastructure across enterprise (private) data centers and public cloud services leads to the most optimal delivery of services, agility, best time-to-market, and cost efficiencies.
5G networks will enable increasing edge data capture and mobility, removing the concerns of data gravity
Enterprises are increasingly deploying applications at the edge; that is, outside of a traditional centralized data center or cloud. That’s not just areas like remote/branch offices but also expanding to non-traditional IT centers like retail locations, airports, sports stadiums and connected/smart vehicles. Many of these locations will likely develop into mini-cloud data centers, requiring full IT stack services with servers, high-speed connectivity and local data storage to deliver these rich new services.
In fact, analysts have predicted that by 2025, three quarters of enterprise-generated data will be created and processed at the edge. The general view has been that mobilizing this edge data will be slow, inefficient and expensive – for instance, once captured at an edge location, data may need to be transferred to a central data center for analysis or other processing. But moving petabytes of data across slower network links is perceived as prohibitive and has given rise to the “data gravity” analogy – that data has “weight” that must be considered before transferring it to another location. While there’s truth in this, we foresee the new generation of 5G network connectivity will have a direct impact on the emergence of data-rich applications deployed on the edge, especially for mobile connected devices. Increasing bandwidth to and from the edge will minimize the concerns of data gravity and further spark the generation and consumption of data on the edge.