Don't underestimate the full stack impact of hyperscale.

It can be easy to think that hyperscale benefits are limited to the infrastructure layer and its management. But the evolution towards software-defined infrastructure impacts every layer of the stack, including infrastructure, composition and resource orchestration, workload execution and data and applications.

As outlined in our new white paper, composition and resource orchestration is an excellent example of this, as the effect of flexible infrastructure can be minimal unless proper decisions are taken at higher levels of the stack.

Ericsson-hyperscale-cloud-infrastructure-composition-resource-composition.jpg

Let's look at three key elements of composition and resource orchestration that are affected when going hyperscale:

Resource scheduling (taking appropriate decisions)

Scheduling mechanisms can be extremely complex in modern datacenters – especially when each individual component (compute, memory, storage, networking and so on) has to be considered.

Due to the limitations of current datacenter architecture, scheduling today involves the selection of the most suitable server to host a particular job, taking into account server properties such as its physical and topological information, as well as datacenter operational policies.

Hardware disaggregation radically changes this process, enabling more fine-grained decisions to be made, more easily taking into account things like power consumption, resource defragmentation, performance and utilization.

Analytics (increasing understanding)

Hyperscale makes prediction through analytics more important than ever and is the cornerstone of any software-defined system that is designed for full automation. Software-defined infrastructure makes analytics more actionable, which better enables infrastructure to be managed in a proactive manner.  With hardware disaggregation, the analytics as well as the resulting actions of that analytics, can be more granular. It could be applied, for example, by identifying and swapping out an individual component that is malfunctioning whereas previously you had to swap out the whole board/server.  

However, events (for example, network-driven or hardware-driven events) cannot always be predicted, so reactive mechanisms – triggered by anomaly detection systems – also need to be in place. Live, accurate information – from the physical infrastructure as well as from the logically composed hosts – must be taken into account by these mechanisms.

Application-aware orchestration (becoming smarter)

Infrastructure is smarter when it knows what it is running. In other words, applications should be described through application profiling of functional and non-functional requirements toward the infrastructure. With hyperscale, the composition and resource orchestration layer has the ability to become smarter at deriving, creating and maintaining the associated infrastructure resources according to an application’s needs.

Reimagining the datacenter

Our new white paper: Hyperscale cloud: reimagining data centers from hardware to applications presents the scope of the impact of a hyperscale datacenter transformation and reveals how the benefits can go beyond the hardware layer to impact the software layer, including hypervisors, operating systems, cloud platforms and – ultimately – applications. Please read the paper and let us know what you think!

Download the white paper


Data & Analytics Cloud Infrastructure

Jesper Tunér

Jesper Tunér joined Ericsson in 2006, after working at Accenture as a management consultant and, prior to that, at a small software company as a programmer. He has held several international positions in marketing, sales and finance throughout Europe, Asia, Africa and the Americas. In 2013 he was appointed Head of the Cloud Marketing Program at Ericsson. Since the start of 2016, he has worked with Cloud Strategy & Portfolio Management for the Product Area Cloud Systems in the Cloud & IP Business Unit. Jesper has a BSc in Computer Science and a MSc in Business Administration from Lund University, Sweden, followed by post graduate courses at IMD Business School and Kellogg School of Management. He is passionate about the digital industrialization opportunities that come with cloud technology.

Jesper Tunér

Discussions