Intelligent Data Centres Issue 12 | Page 63

UNCOVERING THE LAYERS highly distributed nature of modern digital business applications. The key challenges that the combination of Edge Computing and hybrid multi-cloud adoption will solve include: • Lower latency and bandwidth savings: Proximate high-speed, low-latency connections (<60–<20 milliseconds) are necessary for companies to materially close the ‘distance gap’ between their application and data workloads and cloud service providers (CSPs). With agile and scalable cloud environments closer to the users at the Edge, data access and application response times can be faster and cost savings from reduced data transport can be realised. • Enterprise consumption of hybrid multi-cloud: Enterprises generally determine which cloud platform to place their applications on by which CSP delivers the best service for a specific workload. This freedom of choice makes it easy and practical for IT organisations to experiment with different cloud platforms to see which delivers the best quality of service (QoS) at the best price. Additionally, www.intelligentdatacentres.com more than ever before, enterprises require the flexibility of retaining control and securely running business- critical applications in-house and want the flexibility of leveraging both private and public hybrid cloud environments, depending on specific use cases. • Political and regulatory factors: With more frequent and complex incidents of security and privacy breaches, many countries are regulating where and how data can be used. These privacy and data sovereignty compliance requirements will lead to more distributed data centres and cloud services that keep data local to a specific geographic region or country. An aeroplane with thousands of equipment sensors, an autonomous vehicle producing telematics data, or a smart hospital monitoring patients’ well- being can each generate several terabytes of data a day. About 75% of enterprise AI/ analytics applications will use 10 external data sources on average. AI and IoT will drive new interconnection and data processing requirements at the Edge Yet Equinix believes for many use cases, an additional set of stringent requirements related to latency, performance, privacy and security will require that some of the AI/ML data and processing (both inference and model training) be proximate to data creation and consumption sources. Equinix predicts that enterprises will accelerate the adoption of AI and Machine Learning (ML) for a broader set of use cases, requiring increasingly complex and more real-time-sensitive processing of large data sets originating from multiple sources. To meet the scale and agility requirements of the above, Equinix believes businesses will continue to leverage public cloud service providers, while most will likely find ways to use an optimal set of AI/ML capabilities from multiple CSPs-effectively deploying a distributed, hybrid architecture for their AI/ML data processing. Equinix predicts this will create an impetus towards new architectures and the increased adoption of vendor-neutral, richly interconnected, multi-cloud- Issue 12 63