Upgrading data centers to cloud scale efficiency – hype versus fact

“Software-defined” famously made networking sexy, and network efficiency is the New Black. But is anybody wearing it? – If you want to cut through hype and rumour to find out what is really happening, you ask the people at the coal face. That is just what the latest Futuriom report – Untold Secrets of the Efficient Data Center – sponsored by Mellanox Technologies, has done. Over 200 director level or higher data centre professionals were screened by country and company size to dig deeper into actual working practice and the key trends.

“The data centre is being reinvented” according to Scott Raynovich, Chief Analyst, Futuriom. “It’s a real challenge to build a cloud infrastructure that can scale to support demanding applications that can embrace big data, analytics, self-driving cars, and artificial intelligence. The very techniques developed by hyperscale cloud giants are now migrating to the enterprise, where distributed applications now rule. There’s more pressure than ever for networks to perform, and new technologies are beginning to be deployed to make sure that networks don’t become the bottleneck for the cloud. This report provides the most detailed insight into why this matters, and how key players are re-shaping the road map.”


The report summarizes the results of a survey taken in Q1 2019 by Futuriom, and an independent cloud-based data partner. The respondents included 116 from the US, 52 from China, and 50 from the UK – to provide an international overview based on regions where data-centre infrastructure is being deployed aggressively.

By industry, the distribution covered the cloud (49%), telecommunications (26%), and enterprise IT domains (25%). All were screened for IT expertise, with 25% falling into the CxO or SVP category. Roles included: enterprise IT managers (39%), cloud architect/managers (32%), applications development (26%), security (24%), and network manager/architect (22%). The survey was limited to companies with more than 500 employees as follows: 34% of 501-1,000; 41% of 1,001-5,000; 14% of 5001-10,000; and 11% more than 10,000 employees.

The key role of the network

So how is the data centre to be upgraded? Asked to: Rank The Following Technological Responses to Improving Data Center Performance, the highest average ranking goes to: Improve the efficiency of networks using techniques such as processor offload and SmartNICs, whereas the lowest ranking goes to: Deploy more servers.

This theme emerged clearly throughout the survey: that the network is seen as a key engine of performance to the cloud, and it needs specific adaptations to keep up with data centres that have ambitions to be cloud-scale. And the potential benefits expected from these network upgrades include: faster application performance (64%), stronger security (59%), greater flexibility (57%), and application reliability (57%). Overall 84% of respondents thought network infrastructure was either “very important” or “important” to delivering applications such as artificial intelligence and machine learning.

The choice of SmartNICs is interesting as it is a relatively new solution (only 10% confess to not knowing what a SmartNIC is). SmartNICs are Network Interface Cards (NICs) with built-in processors and intelligent accelerators using standard APIs. They can be C-programmed to do anything from optimizing traffic flows to recognising and quarantining malicious data before it reaches a server. This takes an enormous load off the servers that the network connects. Without them, tasks such as Remote Direct Memory Access, Non-Volatile Memory Express over Fabrics (NVMe-oF), compression, encryption, and network virtualization place a constant demand on the server cores, and this reduces the power to support applications. More advanced SmartNICs can even virtualize networked storage to simplify provisioning to both virtual and bare metal servers. basically

So, basically, SmartNICs create a “Smart Network” that manages itself and takes a big load of the servers, freeing them up to provide optimal application support. Asked: Which of the following use cases for SmartNICs appeal to your IT organization?, the overall winner is: Improve efficiency of VMs and/or containers (56%); second is: Virtualize and share flash storage to use it more efficiently (55%). Other choices are: Enable more software-defined networking (54%); Accelerate hyperconverged infrastructure (50%) and Isolate and stop security threats (47%).

There are interesting differences across the three regions polled. For the two most popular use cases ­– improving the efficiency of VMs and containers and virtualizing and sharing flash storage – the responses from China are higher at 65.38% and 75%, respectively. These compare with 55% and 51% in the US, while the UK showed lower levels of interest for all use cases.

The future of Moore’s Law

This overall emphasis on increasing efficiency, as a better way to improve data centre performance than simply adding greater processing power, might be simply a response to the supposed death of Moore’s Law. For decades the IT industry has lived with the comfortable knowledge that processing power would increase and become cheaper year on year. If that is no longer perceived to be true, it might explain this shift from adding hardware to increasing efficiency.

But when asked: Do you believe that Moore’s Law is decelerating, disappearing oe changing how you depend on chip development cycles? The overall majority response (38%) was: Moore’s Law will continue to be around for the foreseeable future, with only 12% saying: Moore’s Law is starting to decelerate or subside. So it does appear that the quest for efficiency marks a positive decision rather than a reaction to reducing opportunity.

Again there were interesting regional variations, with China much more confident about the future of Moore’s Law than the US and UK.


This survey takes a detailed look at the data centre environment and concludes that data centre professionals see the need for new solutions to optimize their operations and efficiency. They want to avoid adding more expensive servers, and they see how virtualization and network optimization technologies offer the best way to achieve their goals.

To that end, they recognise network optimization and SmartNIC technologies as the most realistic way for the average enterprise to upgrade existing data centres to achieve those hyperscale efficiencies.

 | Website

Kevin Deierling has served as Mellanox's VP of marketing since March 2013. Previously he served as VP of technology at Genia Technologies, chief architect at Silver Spring Networks and ran marketing and business development at Spans Logic. Kevin has contributed to multiple technology standards and has over 25 patents in areas including wireless communications, error correction, security, video compression, and DNA sequencing. He is a contributing author of a text on BiCmos design. Kevin holds a BA in Solid State Physics from UC Berkeley.

AI Readiness - Harnessing the Power of Data and AI


Related articles

CIOs and CISOs Battle Cyber Threats, Climate, Compliance

CIOs and CISOs face unrelenting pressure from three massive...

Discover the Power of On-premise Cloud Innovation

For most organisations, the shift from on-premise to the...

The AI Show – Episode 8 – Theo Saville

In episode 8 of the AI Show, our host...

The Data Conundrum: How sustainable is its future?

In this article, Dan Smale, Senior Service Owner of...

Adopting open architecture for robust data strategy

As the world's economy grapples with continuous challenges and...