Hyperscale Computing Demands Hyperscale Innovation

  • Celestica  |
  • 2019-03-18
CloudFest 2019 Hyperscale Blog Post

The rise of the hyperscale data center is placing new pressure on hardware developers to innovate faster in a race to enable breakthrough performance, scalability and reliability within a cost efficient and sustainable construct. Hyperscale emergence is just one of the drivers of a new age of challenges and opportunities we call The Acceleration Economy, where companies must manage the critical issues of innovation, design, production, quality, and productivity head-on and at high speed.  

Driven by big data, the Internet of Things and Cloud computing, hyperscale data centers are causing significant change in enterprise and communications markets. Consider that while a typical data center may support hundreds of physical servers and thousands of virtual machines, a hyperscale data center can support thousands of physical servers and millions of virtual machines. This enables much higher computing power at a lower cost, and the ability to deploy or extend systems faster than ever. 

In the race for hyperscale innovation, those who leverage partnerships to accelerate success will emerge victorious.  A partner like Celestica can accelerate time-to-market by serving as a new source of innovation and investment for the future.  Together with our customers, we leverage our robust hardware portfolio, connections with the software ecosystem, and expert design and engineering team to deliver customized solutions that are optimized for performance.

While hyperscale technology continues to evolve rapidly, the fundamentals are not new, and Celestica has been investing in this area for years. Our robust portfolio of leading-edge, high-performance hardware for enterprise and cloud applications includes next-generation storage, servers and switches, converged racks and hyperconverged enclosures that support the evolution of the data center. 

For example, the Thor 2U Four Node Dense Server Solution is a high performance server optimized for hyperconverged and traditional server applications. It features Intel Xeon® scalable processors and can support pure NVMe or mixed storage technologies. Dual socket server nodes with support for up to 24 DDR4 DIMM sockets, including support for Intel Apache Pass and standard NVDIMMs, provide the power necessary for the most demanding, memory intensive applications. This combination of powerful computing, flexible hybrid storage options and dense memory, sets Thor on the leading edge of emerging hyperconverged solutions.

As companies look for the right partners to enable their hyperscale deployments, one consistent imperative is flexibility. OEMs and ISPs are pushing the envelope with new deployments, leveraging new and different software systems to achieve their goals. Celestica is tightly tied into the vast software ecosystem, giving us the ability to offer tested and certified solutions across multiple software environments. For our customers, the flexibility to leverage their choice of software with certified hardware is a game changer that enables rapid deployment of solutions for hyperscale infrastructure.   

Finally, as the hyperscale market matures and performance demands continue to increase, some hyperscalers have begun to tailor their solutions for a particular application environment - or indeed, even a specific task. We call this customize to optimize.  When deployments evolve to become this tailor-made, companies need a reliable and capable design and engineering partner like Celestica to provide them customized products that optimize performance. 

Operating in the Acceleration Economy means maximizing innovation capabilities, and doing that in the hyperscale environment is no exception. Our switching, storage and server solutions and ability to facilitate custom production and quick delivery make us the optimal hyperscale technology partner.  

Preview some platforms we had on display at CloudFest 2019.