Cohesity CTO: Turning 'backup into heads-up' through generative AI
The Computer Weekly Developer Network sat down this month with Mark Molyneux, EMEA CTO at Cohesity to examine the methodologies, principles and practices that underpin a more proactive information-first approach to backup.
Grasping the opportunities that stem from generative artificial intelligence (gen-AI) and dovetailing this new creative force with the power streams that run enterprise backup has the potential to turn ‘backup into heads-up’ knowledge for actionable business insights, argues Molyneux.
With an almost limitless variety of channels through which businesses can now look to use gen-AI to improve operational efficiencies, what better place than to start at the base-level substrate layer where data itself resides – from this point, we can start to use AI in all its forms (generative, predictive, reactive etc) to unlock the complexity held deep inside enterprise applications, data repositories, services, functions and beyond.
We need to talk, chatbots
From his perspective, Molyneux eyes the role of AI chatbots as a game changer in terms of the way users now interact with increasingly abstracted enterprise technology services. Through use of AI chatbots such as we have seen showcased with ChatGPT, users at any level of technical competency (some without any) can start to transform complex and unwieldy technical data streams into human-readable language that can be quickly interpreted for its business context and purpose.
“This use of gen-AI will not be confined to any particular department or team, so there is a widespread democracy here. But that opening up also comes with risks; organisations that attempt to deploy AI like some band-aid fix or some energy supplement boost with over-ambitious aims can very quickly find that they have created new security gaps and vulnerabilities,” said Molyneux.
As we move forwards into an era of real-world generative intelligence applications within the enterprise stack, Cohesity’s watchwords are AI yes, but AI applied in the most “controlled & responsible” way possible. From an internal development perspective (and with a view to how all AI should be deployed externally), Cohesity champions transparency in order to protect access to the data with role-based access controls.
“This transparency factor goes hand in hand with a wider approach to governance, to ensure the security and privacy of data used by AI models users themselves. This means making sure these data sources and streams are integrated only when they are indexed and classified and securely searchable within a total compute and data environment that is both immutable and resilient,” said Molyneux. “While we often think about generative AI being applied to the user interface surface level, we must also realise how fundamental its impact is as a means of unlocking data (and its business insights) from long-term storage and backups.”
Data protection, security mobility & access
Talking about the need to approach data backup across five central pillars of functionality, Cohesity’s mantra is built upon a foundation of data protection; data security; data mobility; data access; and of course, data insights. Because backup data needs to be protected against the threat of cyberattacks it must also be effortlessly manageable and movable while also keeping it accessible on demand.
“But really that [above statement] is table stakes, it should be par for the course. Our data engineering mission has been to go further than that albeit comprehensive array of capabilities, so that we also provide generative-AI powered tools to unlock hidden data value and be able to bring it forward safely and securely into the operational and transactional layer that an enterprise runs with everyday,” said Molyneux.
Cohesity EMEA CTO Mark Molyneux:: On a data engineering mission to ‘go further’ and now harness gen-AI to go from back-ups to heads-ups!
Looking at the company’s products and services in this arena, Cohesity Gaia is the company’s first intersection point that combines the intelligence derived from Large Language Models (LLMs) and the data repository tools and services that we know it for. With this new marriage of unified technologies, Cohesity says it can drive conversational search, and question-and-answer services which also feed into knowledge management, together with the valuable context by providing the information source, avoiding hallucination concerns.
Deep inside data streams
Molyneux rounds out this discussion by explaining how, today, users will be able to create a semantically aware index of data that enables software applications and services to extract, manage and channel valuable nuggets of source material from deep inside data streams or repositories that may have otherwise been passed by, consigned to ignominy, regarded as candidates for data lake obscurity or otherwise left comparatively unobserved and unloved.
“By identifying the data that is a point of action via what is quite simply natural language searches and prompts – without the need for complex code or query languages – we can promote an enterprise’s backup infrastructure to an elevated position where it significantly impacts (and improves) operational efficiency and the corporate bottom line,” concluded Molyneux. “This is a unique opportunity to make proactive and progressive use and reuse of backup data while safeguarding proprietary information from accidental exposure with appropriate levels of user access control.”
The shift from data backup to business opportunity (or operational efficiency) heads-up may represent a solid foundation block and new ray of light for a whole spectrum of generative AI functions that now span from the data centre to the digital toaster.