Over the last few years, innovations by the cloud giants have disrupted many traditional industries. Amazon and Alibaba changed the way we shop, Netflix and YouTube brought us new ways to consume video, and WeChat and Facebook revolutionized how we connect and communicate with our friends and families.
The agile, elastic, and highly efficient cloud infrastructure is one of the key pillars of innovation these cloud giants are delivering. As a result, they have become the “North Stars” of efficient service delivery and data center architecture. Organizations look to Google, Facebook and others for cloud and data center infrastructure best practices when they set out to build their own to support their innovations and business transformation.
So what are these top cloud service providers (CSPs) doing to modernize their cloud infrastructure now?
They were eager to get their hands on the new Intel® Xeon® Scalable processors as soon as possible. I couldn’t be more excited to see them deploy Xeon® Scalable, modernize their infrastructure, and create new, differentiated services. More than 500,000 Xeon® Scalable processors have already shipped to over 30 customers, many of them CSPs. Based on my conversations with them, here are three top reasons CSPs are so thrilled about Xeon® Scalable.
1) Disruptive performance efficiency across cloud workloads
For CSPs, the math is pretty simple when it comes to profitable growth of their business. Because they sell computing as a service, price performance of their infrastructure, often measured with performance/total cost of ownership (TCO) dictates how efficiently they can generate profits from their services. When a new breed of processor such as Xeon® Scalable offers a performance boost that overwhelms the cost of upgrading, and a 65% lower TCO in a typical virtualization deployment scenario compared to a 4-year-old server**, CSPs jump on it.
Google launched Xeon® Scalable-based cloud instances on Google Cloud Platform (GCP) in February of this year and made the instance generally available on May 31. Bart Sano, Google’s Vice President of Platforms, told us that GCP customers running applications such as bioinformatics, machine learning on satellite imagery and finance simulation and report up to 40% improvement in many cases compared to the previous platforms. In some cases, where the customers tuned their software for Intel® Advanced Vector Extensions 512 (Intel® AVX-512), they saw more than 100% improvement.
Amazon Web Services, announced its new C5 instance last fall and called it the most powerful compute optimized instance and the best price to compute performance in EC2. This summer, Dr. Matt Wood, Amazon’s General Manager of Artificial Intelligence and an iconic figure in the AI space, stated that the computational power of the Intel® Xeon® Scalable processor lets AWS customers use more data, to create innovative new products and experiences powered by machine learning, increasing machine learning inference performance by over 100 times.
2) Enable differentiated services to expand into new markets
New and emerging use cases such as AI, autonomous driving, personalized medicine, media and graphics, and gaming present opportunities for CSPs to innovate and monetize new user experiences. The workloads supporting these new services are normally compute or data intensive, and traditionally run on purpose-built hardware clusters. With the workload-optimized performance and acceleration Xeon® Scalable Platform brings, it is now feasible to offer these services in the cloud
First, the high-performing cores of Xeon® Scalable can unearth new opportunities. Features such as Intel® Advanced Vector Extensions 512 (Intel® AVX-512) can accelerate performance by a factor of up to 2x over the previous generation of a range of compute-intensive applications, including security/crypto/compression, visualization, in-memory databases, deep learning, and modeling and simulation in HPC. 1
Genomic sequencing and analysis used to mainly run on supercomputers, making is challenging for researchers worldwide to access and share resources. Seven Bridges, a biotech startup, now provides a biomedical data analysis platform as a differentiated service in healthcare vertical on Google’s GCP to accelerate cancer diagnosis and precision treatment. With another 1.8X performance boost on Xeon® Scalable instance, Seven Bridges CEO Dr. Brandi Davis-Dusenbery believes its customers “will benefit from significant efficiencies and can use more of the world's biomedical data to accelerate discoveries and improve health.” 2
In addition, Xeon® Scalable goes beyond a processor, it is a flexible platform that can integrate with accelerators programmed to optimize the performance of emerging workloads, enabling a slew of new possibilities. Accelerators like Intel® FPGA give the cloud the speed, predictability, and energy efficiency along with easy customizability to take on emerging workloads such as AI, HPC, advanced analytics, and IoT, while the Intel® NervanaTM suite of products can accelerate AI training.
Microsoft Azure based its Configurable Cloud on Intel® Xeon® and Intel® FPGA to achieve a degree of flexibility and performance that makes it much easier to create new cloud services to meet the emerging demands AI and analytics that are powering the next generation of enterprises. Kushagra Vaid, GM of Azure Hardware Infrastructure, Microsoft’s Azure Platform, also anticipates that Intel accelerators, such as Intel® FPGA, Intel® Xeon PhiTM and Intel® NervanaTM all will play a key role in this next generation of infrastructure to help the application developers be more productive in creating new services and new user experiences.
3) Joint drive toward cloud innovation
The workloads that the CSPs run on top of Intel® architecture are different, and the cloud services they create are diversified, but there is one thing in common among the feedback we got from the top CSPs. Every CSP collaborated with Intel one way or another to jointly drive new innovations in the cloud space. For some CSPs, such as Amazon and Google, the collaborative engineering efforts extend back a decade ago. From virtualization to containers, from a relational database to NoSQL, big data, AI and machine learning, Intel has been on the forefront of cloud computing and collaborating with CSPs to drive cloud innovations for years to make sure cloud workloads run best on Intel Architecture - faster, more efficient and more secure.
More than 25 CSPs showcased their cloud services based on Intel® Xeon® Scalable processors at our launch last month, and it is not too late to check them out. Visit our Xeon® Scalable portal to find out how some of the most successful CSPs are putting Intel® Xeon® Scalable processors to work.
1 Up to 2X flops per cycle: when compared with Intel® AVX2 available on previous E5/E7 v3, v4 families
2 1.8X faster than previous generation: when comparing Seven Bridges Graph Suite's end-to-end whole genome processing pipeline running over Xeon® Scalable processor based GCP instances to previous E5-based GCP instances.
Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations, and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more complete information visit http://www.intel.com/performance.
** Up to 65% lower 4-year TCO estimate example based on equivalent rack performance using VMware ESXi virtualized consolidation workload comparing 20 installed 2-socket servers with Intel Xeon processor E5-2690 (formerly “Sandy Bridge-EP”) running VMware ESXi 6.0 GA using Guest OS RHEL6.4 compared at a total cost of $919,362 to 5 new Intel® Xeon® Platinum 8180 (Skylake) running VMware ESXi6.0 U3 GA using Guest OS RHEL 6 64bit at a total cost of $320,879 including basic acquisition. Server pricing assumptions based on current OEM retail published pricing for 2-socket server with Broadwell based Intel Xeon processor systems– subject to change based on actual pricing of systems offered.