Questions Your CIO and CDO Will Ask About Multi-Cloud and Hybrid-Cloud (Guest blog from SAS)

Author: David Shannon, Head of Hyperautomation, SAS Software Ltd.

Today, many large-scale organisations operate in a multi or hybrid cloud environment.  Orchestrating a mix of cloud and data centre infrastructure is a growing challenge.  Despite the benefits of flexibility, scalability, security and avoiding a single point of failure, data sprawl, lack of interoperability, and under-utilized features are pushing up costs for human capital.

Research by SAS found 99% of technology professionals in large companies are experiencing difficulties in multi and hybrid cloud environments.  This is compromising the accuracy and timeliness of analytics; leading to an increased risk of poor decision-making across the organisation.

Over 300 organisations have contributed to research and roundtables during 2023.  Here are three sets of questions your CDO and CIO are likely to be considering about muti- and hybrid cloud environments.

Chief Data Officer

The CDO, or individual with the de facto role, is primarily concerned with supporting fact-based decision making, innovation, and so on, to support delivery of the organisation’s strategy.  This group are concerned with value, accuracy and skills. 

1. What value is returned?

Understanding the intended success criteria will ensure value is be measurable.  Value can be attributed in different ways.  For example:  Speed of computation means faster, more efficient analytics leading to fewer cloud computation resources.  This directly reduces operational costs and increases workforce productivity2.  Contributions to social charters can be included, through reduced power consumption.

Consider platforms that demonstrate benchmarked scalability and optimised computation for AI and Analytics.

2. How will we maintain accuracy?

SAS research1 found that multiple clouds not only increase costs, but also hide answers.  Data is commonly duplicated between multiple environments due to lack of seamless, secure integration. 

Multiple environments lead to analyses being performed in multiple locations and inevitable discrepancies between results, or data being manually transferred between environments.  In each case this causes data quality issues such as latency of data availability, leading to incomplete data or inaccuracy.  This compromises the effectiveness of analytics which in turn leads to an erosion of trust. 

Organisations are increasingly acting on trustworthiness.  Artificial intelligence offers increasingly powerful methods of insight and automating processes.  Fairly and accurately represent your subjects in decision making with AI is the concern of shareholders, regulators and governments. 

3. What skills are required?

STEM skills shortages are a persistent challenge, hence introducing another data or analytics platform leads to further skills requirements and an increased dependency on the successful passage of knowledge in the medium and long term.

Analytics should be approachable irrespective of skills.  Consider why your organisation uses programming languages at all?  Could your analytics requirements be delivered with no-code and low-code applications?  In reality, there will be at use cases where the versatility of a scripting language is required.  However, the language should be considered simply a that, a means to communicate with the same underlying platform which derives analytics from a common data source with a common set of validated routines.

Consider multi-modal platforms that allows repeatable analytics to created and governed.  The platform should support a bring your own language approach, supporting openness without data spawl or unvalidated algorithms. 

Chief Information Officer

The CIO is concerned with the IT systems which run and operate data and analytics platforms.  Amongst other points, this group is concerned with security, costs and data sprawl.

1. Will it adhere to our Security strategy?

With systems between clouds or split between a cloud and legacy data centre, integrated authentication and authorisation is not always possible. 

A legacy system may hold valuable data for reporting, model training, forecasting and more; but accessing it from the cloud depends on local authentication and authorisation.

Consider an analytics platform which provides a data agent running inside the data centre.  This should facilitate access to any data platform, and provide secure connectivity to the Cloud and use local database authentication and authorisation systems.

2. How will we balance Cost and Innovation?

SAS research showed that around one third of cloud expenditure is wasted1.  This is caused by a lack of visibility and cost tracking becoming too onerous.  The CIO’s office is increasingly prioritising cost management. 

One set of efficiencies can be gained through push-down analytics, and running advanced analytics in situ.  This minimises movement of data, both benefiting security, omitting duplication of data, the need for data silos and egress costs. 

Consider analytics platforms that support techniques for push-down analytics and embedded processing.  Consider platforms that provide open transparent usage metrics allowing targeted process improvement.

3. Yet another analytics platform?

Organisations seek a single platform that covers the entire analytics life cycle.  For those organisation yet to discover SAS Viya, it is understandable that multiple solutions are procured to address different needs.  This contributes to cloud sprawl and challenges maintaining a lineage and a catalogue of assets used to create and run analytics.  In turn this leads to further challenges of Observability.  Similarly to the CDO’s office who need to demonstrate trustworthiness of analytics, the CIO’s office needs to demonstrate integrity and security, specifically what data is accessed, how often, by what or who? 

The capability to automatically build a data catalogue, describing its contents with schematics and a business glossary should be selected.  This aids reuse of data and informs data literacy within the organisation. 

Summary

For the multi- and hybrid cloud scenario, we have considered the challenges of CIO and CDO offices separately.  This is not to create division, but demonstrate that different viewpoints in are necessary and valid.  CIO and CDO offices work together to achieve valuable, efficient and performant, secure and governed platforms.  

References

1. Multi Cloud and Analytics Report | SAS UK, 2023.
2. SAS Viya outperforms competitors and lowers operational costs, study says | SAS UK

---