The telecom community is suddenly embracing open hardware designs, specifically those developed in the Open Data Center Committee (ODCC) and the Open Compute Project (OCP). The natural reaction is why and why now? Let’s take a look at some answers.

Historical background

The OCP (Open Compute Project) was originally started by Facebook, working together with Quanta. They wanted to drive standards on how Facebook delivers hardware in its own hyperscale datacenters. Facebook had been inspired by how great ODM firms are at taking a standard reference design and then continually evolving a platform – much like ODMs do with Intel x86 designs. So Facebook decided it would be better to actually open source its own hardware design to both speed up innovation and drive down cost over time.

About the same time, some of the biggest Chinese web companies started a similar initiative, called Project Scorpio, with the same motivation – to drive down capex while having closer engagement and involvement with the hardware design and vendors. The initial focus ofProject Scorpio was to drive standards across mechanical and rack level design in shared power and shared cooling. But it then evolved into datacenter design and deployment phase and was rebranded as the Open Data Center Committee.

For those not familiar with how big these Chinese companies are, Baidu is the dominant search engine in China (think Google), Alibaba is the dominant e-commerce platform (think Amazon, PayPal and eBay) and Tencent is the internet landlord that all Chinese internet firms farm on and also provides the widest used mobile messaging and payment platform, WeChat (think Facebook, Instagram and Skype). The other founding members are China Mobile and China Telecom and the China Academy of Information and Telecommunication Technology (CAICT).

Why is ODCC important?

From a market perspective, China has the fastest growing market in mobile usage and datacenter growth rate, and, as you can see above, ODCC involves the largest Chinese web and telecom companies. The standardization of hardware design/deployment/operations across the Chinese web community and the telecom community is going to have huge implications in the way we all deploy applications, utilize the network and go from centralized to distributed cloud infrastructure.

Much like mobile networks today, we need to establish certain infrastructure standards so top layer efficiencies can be utilized. Today we take for granted that we can take our mobile phones to any corner of the world and just expect them to work regardless of the mobile network provider. But this took years of evolution and standardization efforts to accomplish, and the web developers of today already expect that same experience – to just package up an application and deploy it in any corner in the world, whether it’s running a drone, a car, or enabling a smart device.

Why join both ODCC and OCP?

Both ODCC and OCP are projects driven by the biggest players in the world’s largest web markets, China and the US. Both are trying to standardize across datacenter deployment and innovationin hyperscale. One of ODCC’s missions is to drive modular datacenter designs and converge the Scorpio rack deployment. During Mobile World Congress, Ericsson announced that it was joining ODCC and OCP. By being an active member in both groups – and a Platinum member of OCP – Ericsson is committed to driving the standardization effort across both communities as well as both capital and operational efficiencies.

What does the next generation datacenter mean to CIOs?

Datacenter assets are becoming increasingly strategic, which means they will be viewed through a TCO lens, not the traditional capex/opex view. For CIOs, the question of how to manage an efficient datacenter and simplify the underlying vendor islands are critical issues in reducing opex and moving from traditional OEMs to a factory direct model, while at the same time avoiding supply chain and logistics issues.

If he application-centric web companies of the world work together with network-centric telecom operators, they can build a common underlying infrastructure. This will allow CIOs to be more agile in workload movements, see better TTM due to simplified operations and know that their datacenter TCO will provide economic metrics to their board.

Leading technology in software defined infrastructure

We are driving the standards and deployments for next generation datacenters with our primary customers – telecom operators – and the web giants of the world. We have the world’s leading technology in software defined infrastructure, solutions, and services and experience in building the world’s largest distributed infrastructure running and operating mobile networks.

If you want to learn more about our work with OPC and ODCC is furthering our vision of a Networked Society driven by cloud, 5G and the IoT, please check out our recent announcements on ODCC and OCP respectively.

To learn more about Ericsson's announcements in hyperscale IT infrastructure and cloud, replay our webinar:

Hyperscale now webinar on demand


Partnerships Cloud Infrastructure

Howard Wu

Howard Wu is the Head of Product Line Software Defined Infrastructure for Business Unit IT & Cloud Products at Ericsson. Howard joined Ericsson in 2014. He has a long standing career in technology having worked from hardware manufacturing to software services in countries from China, Canada, US and Sweden. Howard founded Layerboom in 2009 and was acquired by Joyent, a US leading cloud service provider in 2010. Since the acquisition, Howard acted as General Manager for Joyent Canada and later President of Greater China for Joyent prior to joining Ericsson. Throughout his career, Howard has driven successful outcomes through strong team cultures built around collaboration, understanding and meeting business needs while continuing to drive the best in class technology forward.

Howard Wu

Discussions