What is Hyperscale Computing?

Hyperscale computing describes a flexible data center architecture that can be rapidly scaled on demand using large horizontal server arrays and software defined networking (SDN). Specialized load balancing software directs traffic between clients and servers.

• Hyperscale computing provides unprecedented levels of data throughput and hardware efficiency. Artificial intelligence (AI) is used to optimize the computing, networking, and big data storage processes, tailoring them to evolving service requirements.

• Virtual machine (VM) or containerization processes allow software applications to be moved easily from one location to another. Container instances can also be replicated easily when the demand increases.

• The capacity and flexibility provided by hyperscale technology makes it a popular option for cloud computing and big data storage. Both public and private cloud environments utilize hyperscale computing architecture.

• The adoption of Machine Learning (ML) and the Internet of Things (IoT) will further accelerate the growth of hyperscale computing companies.

Cloud computing and hyperscale computing are similar and overlapping concepts. In each case, offsite computing and storage resources are used to increase customer capacity quickly without adding internal infrastructure or IT support.

• Cloud Computing is defined as the delivery of computing services over the internet. This includes computer power, applications, and data storage. Access to these services on a pay-as-you-go basis allows businesses to scale appropriately and lower their overall OpEx.

• Hyperscale Computing is characterized by the scalability of the computing resources and the size of the data center. A minimum threshold of five thousand servers on a ten thousand square feet or larger footprint, as defined by the IDC, is accompanied by a discrete software layer and advanced automation to free applications and services from hardware constraints.

• Hyperscale Cloud Computing brings the best of both concepts together to ensure cloud offerings like infrastructure as a service (IaaS) and software as a service (SaaS) can be scaled up quickly to meet increasing demand. Public cloud providers commonly operate in the hyperscale category.

Hyperscale computing companies utilize the latest hardware and software technology to ensure a high level of reliability and responsiveness to customer demand. As virtual monitoring solutions and the IoT lead to more unmanned hyperscale deployments, visibility will improve as CO₂ emissions decline.

• Unlimited Scalability: The scalable architecture of hyperscale computing can accommodate peak demand levels. When data centers do approach their capacity limits, distributed computing, made possible by high-speed data center interconnects (DCIs), seamlessly extends the network geographically to tap into available resources.

• Efficiency: Automation, software defined networking (SDN), and UPS power distribution methods help to reduce overall energy consumption. Custom airflow handling and balanced workload distribution across servers optimize cooling efficiency. Industrial IoT (IIoT) temperature and power sensors add additional layers of intelligence and efficiency to the feedback loop.

• Redundancy: By utilizing containerized workloads that can easily be migrated between servers, hyperscale computing companies maintain redundancy without significantly increasing their power output. Important applications and data are preserved in the event of an outage or breach.

Disadvantages of Hyperscale Computing

The scale associated with hyperscale cloud computing provides benefits and performance levels that cannot be attained by a conventional data center. At the same time, high traffic volumes and complex flows can make real-time hyperscale monitoring difficult. Visibility into external traffic can also be complicated by the speed and quantity of fiber and connections.

• Security issues are magnified by the size of the hyperscale data center. Although proactive security systems are an essential part of cloud computing, a single breach can expose huge amounts of sensitive customer data.

• Construction schedules are compressed by the increased demand for internet content, big data storage, and telecom applications. These pressures can lead to minimized or omitted pre-deployment fiber and performance testing. Automated fiber certification and network traffic emulation tools minimize schedule impact while significantly reducing post-deployment service degradation.

• Hyperscale data center proportions will continue to expand, even as the available talent pool continues to shrink, especially in remote or undeveloped regions. More automation, machine learning, and virtualization is needed to prevent the demand for resources from overwhelming the ecosystem.

• Greenhouse emissions are a growing concern for hyperscale computing companies. Data centers already consume approximately 3% of the world’s electricity. This has prompted many leading cloud computing companies and data center owners, including Google and Amazon, to pledge carbon neutrality by 2030.

What is Hyperscale Cloud Computing?

Hyperscale cloud computing is defined as the provisioning of hyperscale architecture, including advanced load balancing, horizontally networked servers, and virtualization, to provide scalable computing services and applications to customers over the internet. The features of cloud computing by Hyperscalers continue to evolve as new use cases and technologies develop.

• 5G Adoption is driving more network disaggregation and distributed architecture into the cloud computing model. Despite the emphasis on intelligent edge computing to support the IoT and low latency 5G use cases, hyperscale cloud computing will continue to support 5G Core functionality, big data storage, and the artificial intelligence required for network slice orchestration.

• Automation is essential for maintaining the performance levels of increasingly virtualized and decentralized networks. Improved automation reduces operating costs for hyperscale computing companies by optimizing server and cooling system power consumption. Automation is also an important element of hyperscale data center test practices including DCI fiber certification and monitoring.

The Future of Hyperscale Computing

Although it is impossible to predict the future shape and direction of hyperscale computing, it is certain that the unprecedented demand for computing services and big data storage will continue unabated. Fueled by the advent of 5G, the IoT, and artificial intelligence, this data center market size is expected to multiply in the coming decades.

These market factors are also moving hyperscale cloud computing into a more distributed model. Edge computing will accelerate continually, slowly shifting intelligence and storage closer to the expanding universe of IoT sensors and devices. Complex fiber networks connecting these locations will heighten the importance of automated pre-deployment MPO-based fiber testing and high-speed transport testing.

The reality of unmanned hyperscale data centers will contribute to a sharply reduced carbon footprint. Many leading hyperscale computing companies have already committed to 100% renewable energy sources. Innovative projects like Microsoft’s Project Natick subsea data center and Green Data ground and rooftop solar panels prove that hyperscale computing can be re-imagined to sustainably coexist with the environment.   

VIAVI has established a suite of end-to-end (e2e) test and monitoring solutions designed to support the evolving hyperscale computing ecosystem. Automated test workflows and unmatched expertise extend from initial testing and proof of concept in the lab through comprehensive maintenance and troubleshooting in the field.

• High-Speed Transport - 400G and 800G: As cloud architectures becomes more distributed and edge computing expands, high-speed Ethernet standards like 400G and 800G are essential for maintaining the capacity and latency requirements of hyperscale computing companies. Y.1564 and RFC 2544 service activation tests assess latency, jitter, and frame loss. Automated test solutions also verify ongoing performance through advanced error rate and throughput testing.

• Virtual Test: Service activation and performance testing that once required a technician on site can now be performed virtually using the innovative VIAVI Fusion test agent, optimized for cloud-native architecture. Ethernet activation tests including RFC 6349 TCP throughput testing can be virtually initiated and executed. Network performance metrics and SLAs are verified confidently for even the largest service providers.

• Observability and Validation: Machine learning and artificial intelligence are moving hyperscale cloud computing to the next level by allowing complex decision-making and traffic management functions to be automated. The visibility and actionable intelligence provided by these innovations also extends to VIAVI hyperscale test and monitoring solutions.

  • Validation is critical for new services designed to run over virtualized cloud platforms. The TeraVM software appliance is ideal for validating network components in the lab or predicting the performance of virtualized applications. This versatile tool also emulates RAN and Core elements and common internet threats to validate wireless network features and security applications.

  • The Observer network performance monitoring and diagnostics platform combines wire data with enriched flow records to establish a 360° view of network health and utilization. Intuitive, end user experience scoring provides hyperscale computing companies with insight into traffic flow and performance from an individual customer perspective.
    

    

As cloudification blurs the lines between networks and applications, and 5G adoption pushes intelligence to the edge, hyperscale computing is being completely redefined. Industry-leading test solutions from VIAVI help operators plan, deploy, troubleshoot, and optimize networks and services throughout their lifecycle.

VIAVI has leveraged decades of experience and industry collaboration to deliver a suite of automated, cloud-enabled test solutions that support the hyperscale computing ecosystem from end-to-end. This comprehensive approach extends from use case emulation and validation in the lab to fiber certification, high-speed Ethernet transport testing, and remote network monitoring in the field.

As an active participant in over thirty standards bodies, VIAVI continues to anticipate and influence hyperscale trends to stay ahead of the technology curve. Unwavering commitments to open APIs and test process automation (TPA) allow hyperscale computing companies to write their own automation code while keeping up with the escalating pace and scale of hyperscale deployment.