Digital Transformation, Enterprise Solutions, and My DASH Approach
Over the past several years, while assisting many customers, particularly following my attendance at the NRF ’24 Retail’s Big Show and my presentation there, one question has lingered in my mind: what are the most effective systems, solutions, and tools for digital transformation? A simple Google search yields numerous blogs promoting specific systems, platforms, or tools. However, we all know that there’s no one-size-fits-all architecture or solution.
As I prepare materials for a conference presentation on “Digital Transformation and AI”, I have come to realize that while there may not be a one-size-fits-all solution for every business scenario, there are fundamental considerations that every company must address to find the solution that best suits their needs — my DASH approach.
My DASH approach
In light of current market trends and technological advancements, when evaluating a solution, system, or platform for digital transformation, or planning to develop one in-house, I recommend a:
- Data oriented,
- AI/ML enriched,
- Secured
- Hybrid cloud
solution.
D as data oriented and data driven
I have published several blogs about the significance of “Data Oriented Computing” and architectures before. These posts emphasize the crucial role of prioritizing data as the central consideration for any IT solution:
- IT to business is about turning data into value
- Data investments accelerate business growth
- Data is a central element of business innovation
- AI is only as good as your data. However, current trend is there are more data in more locations with more formats, but less quality.
Data collection, quality assurance, organization, analysis, integration, and consistency are all foundational considerations for an enterprise solution.
A as AI/ML enriched
With the advent of ChatGPT 3.5, AI/ML is becoming the next technological revolution. AI/ML technologies encompass a broad spectrum, ranging from hardware innovations (such as GPU, quantum computing, and photonic computing), to advanced models (including transformer, diffusion, and the recent Mamba model), as well as techniques like fine tuning LLM/GenAI, prompt engineering (including RAG), and AI agents and tools.
LLM/GenAI bridges the gap between business and IT by serving as a conduit between business cases and technological solutions. Effectively leveraging AI/ML models and technologies can significantly benefit businesses, as highlighted by Gartner’s report below:
Just last week, Stanford’s Institute for Human-Centered Artificial Intelligence published AI index Report 2024, including following topics:
- AI beats humans on some tasks, but not on all.
- Industry continues to dominate frontier AI research.
- Frontier models get way more expensive.
- The United States leads China, the EU, and the U.K. as the leading source of top AI models.
- Robust and standardized evaluations for LLM responsibility are seriously lacking.
- Generative AI investment skyrockets.
- The data is in: AI makes workers more productive and leads to higher quality work.
- Scientific progress accelerates even further, thanks to AI.
- The number of AI regulations in the United States sharply increases.
- People across the globe are more cognizant of AI’s potential impact — and more nervous.
S as secured
No IT solution can be considered secure without robust security measures in place. With the widespread use of data and AI/ML technologies, IT security encompasses a broader scope, including:
User secured:
- User authorization and
- User authentication
Data secured:
- Data governance
- Data quality
- Data consistency
- Data leaking
AI/ML secured - TRiSM (trust, risk and security management):
Current AI/ML technologies have following issues:
- Model hallucination
- Model bias, fairness and drift
- Model lack of explainability
- Unintended loss of intellectual property
- Legal risks and potential for misuse
GenAI makes the solutions security more complicated, as:
- LLMs generate new content based on learned information.
- LLMs may leak data due to the model’s memorization and overfitting on training data.
Policies and Regulatory:
Like Stanford AI Index Report mentioned, the number of AI-related regulations has risen significantly in the past year. For example:
- President Biden Issued Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence last October.
- EU AI Act published last August
Enterprise solution also need to put these into consideration.
So, to have a secured IT solution, following are my suggestions:
- Determine use cases and restrictions
- Fully understand your data
- Understand GenAI current capabilities, strengths, limitations, and risks
- Evaluate the solution platform (hybrid, public, private) and security
- Establish risk tolerance
- Agree on decision rights and risk ownership
- Determine disclosures
- Alignment among stakeholders
- Reassess the organizations’ existing governance/audit frameworks.
H as hybrid cloud platform
Cloud native computing is an approach in software development that utilizes cloud computing to build and run scalable applications in modern, dynamic environments such as public, private, multicloud, and hybrid clouds. (Wikipedia)
To ensure that an IT solution remains composable and flexible enough to adapt to rapidly evolving technologies, a hybrid cloud platform is ideal. A hybrid cloud platform is a cloud computing environment comprised of on-premises, private, and public cloud services, designed with an open ecosystem in mind. These services can be sourced from various service providers.
Hybrid cloud architecture:
Hybrid cloud solution will include layers like above image shown, and put following into consideration:
- Cloud agnostic
- Consistent management and operation behavior
- Seamlessly AI/ML integration
- Real-time data flow and integration
- DevOps/MLOps
- Open ecosystem
- Functional integration among Cloud, IoT, and Edge.
The design of a hybrid cloud platform enables the IT solution to be flexible and adaptable to future market demands, as discussed in my previous blog MVP vs MVP design.
Furthermore, as computing evolves with advancements such as CPU vs. GPU and classic computing vs. quantum computing, hybrid computing is poised to become the natural choice for the future.
Hybrid cloud operations
Hybrid cloud operation automation is another key feature of hybrid cloud platforms. Platforms like Red Hat Openshift Container Platform and Ansible Automation Platform offer robust functional support in this regard.
Red Hat blog “The importance of automating the hybrid cloud” had some good discussion.
I also like the MLOps Eternal Knot diagram posted with blog MLOps: Machine Learning Lifecycle. I extended the knob to meet with my DASH platform:
Key Takeaways
In any IT solution, data always serves as the central component — retrieved from various sources and transformed into result data that is valuable to the business and consumable by end-users.
AI/ML technologies have the potential to assist with data analysis, operational tasks, natural language processing, and significantly reduce human effort.
However, the behavior of AI/ML models underscores the critical importance of solution control, governance, data and AI model trust, risk and security management.
To ensure the sustainability and adaptability of a solution to evolving market and technological landscapes, leveraging a hybrid cloud platform provides the best opportunity for an open ecosystem within the IT infrastructure.
This is my DASH approach.
