Network configuration error blamed for AT&T wireless outage

The Federal Communications Commission (FCC) has issued a comprehensive report detailing the causes and impacts of a nationwide AT&T wireless service outage on February 22, 2024. The outage, which lasted over 12 hours, prevented customers from using voice and data services, blocking more than 92 million phone calls and over 25,000 attempts to reach 911. The report also includes recommendations to prevent similar incidents in the future.

FCC Chairwoman Jessica Rosenworcel emphasized the gravity of the situation, stating, “When you sign up for wireless service, you expect it to be available when you need it – especially for emergencies.” The outage not only disrupted consumer communications but also affected public safety personnel using FirstNet. The FCC’s Public Safety and Homeland Security Bureau promptly launched an investigation, revealing several key findings about the outage’s extensive impact and the subsequent corrective actions taken by AT&T.

The outage affected users in all 50 states, Washington, D.C., Puerto Rico, and the U.S. Virgin Islands, impacting over 125 million devices. During the critical early hours of the outage, all 4G voice and 5G data services were unavailable to AT&T customers, including FirstNet subscribers. This disruption particularly impacted FirstNet, as device registrations approached normal only after the restoration of the dedicated network elements connected to AT&T’s broader network.

The FCC report highlights the technical cause of the outage, pinpointing two key steps that led to the network configuration error. An incorrect configuration was first made by an AT&T employee, followed by another employee loading this erroneous network change. This sequence revealed insufficient oversight and controls within AT&T’s processes, allowing the misconfiguration to propagate. Additionally, the network’s inability to handle the sudden influx of re-registration requests once the error was corrected prolonged the outage significantly.

Key Points:

• The outage blocked over 92 million voice calls and prevented more than 25,000 calls to 911.
• AT&T prioritized restoring FirstNet services but delayed notifying FirstNet users about the outage.
• The outage was caused by a network configuration error due to insufficient oversight and controls.
• The downstream network element lacked controls to mitigate the error, triggering Protection Mode and disconnecting all users.
• System limitations caused registration congestion, prolonging the outage even after the initial error was corrected.
• AT&T has since implemented additional technical controls, forensic work, and peer review procedures to prevent similar issues.

Corrective Actions by AT&T:

• Implemented additional technical controls within 48 hours of the outage.
• Scanned and updated network elements to prevent similar errors.
• Enhanced network robustness and resilience through ongoing forensic work.
• Adopted new procedures ensuring maintenance work requires completed peer reviews.
• Improved registration systems to handle higher capacity and quicker recovery from Protection Mode.

The full FCC report can be accessed here..

Read More

Samsung Electro-Mechanics partners with AMD for substrates

Samsung Electro-Mechanics (SEMCO) has announced a collaboration with AMD to supply high-performance substrates designed for hyperscale data center compute applications. These advanced substrates are produced at SEMCO’s technology hub in Busan and a newly constructed state-of-the-art factory in Vietnam. SEMCO has invested 1.9 trillion KRW (approximately $1.6 billion USD) in the new FCBGA factory, highlighting its dedication to enhancing substrate technology and manufacturing capabilities to meet the highest industry standards and future technological needs.

The partnership with AMD aims to address the challenges of integrating multiple semiconductor chips (chiplets) on a single large substrate. These substrates, crucial for CPU and GPU applications, provide larger surface areas and higher layer counts necessary for dense interconnections in advanced data centers. Compared to standard computer substrates, those used in data centers are ten times larger and feature three times more layers, ensuring efficient power delivery and signal integrity. SEMCO’s innovative manufacturing processes minimize warpage, ensuring high yields during chip mounting. The FCBGA factory is equipped with real-time data collection and modeling capabilities, allowing SEMCO to develop predictive manufacturing models that ensure the integrity of signal, power, and mechanical components, positioning SEMCO as a leader in next-generation data center substrates.

Key Points:

• Collaboration with AMD focuses on hyperscale data center compute applications.

• Substrates are produced in Busan, South Korea, and a new factory in Vietnam.

• SEMCO invested 1.9 trillion KRW (approximately $1.6 billion USD) in the FCBGA factory.

• Substrates support CPU/GPU applications with larger surface areas and higher layer counts.

• Data center substrates are ten times larger and have three times more layers than standard substrates.

• Advanced manufacturing processes ensure efficient power delivery and signal integrity.

• The new factory features real-time data collection and predictive modeling capabilities.

• SEMCO leads in producing substrates with passive and active components for next-gen data centers.

Read More

NTT and Okayama University show gigahertz ultrasonic circuit

NTT and Okayama University have developed a gigahertz ultrasonic circuit that utilizes the principle of topology. This innovative development promises to revolutionize the design and performance of ultrasonic filters used in smartphones and IoT devices.

The new technology employs a special material called ultrasonic topological phononic crystal, created using an artificial elastic structure with periodic arrays of microscopic holes. This unique design allows ultrasonic waves to propagate stably through complex, microscopic channels without backscattering, a problem that plagues conventional ultrasonic circuits.

The implications of this breakthrough are significant for the wireless communication industry. By solving the issue of reflection in folded small waveguide structures, the researchers have paved the way for the miniaturization and enhanced performance of ultrasonic filters. This advancement has applications for the evolving needs of 5G networks and the growing IoT ecosystem, where efficient signal processing and minimal interference are paramount.

The research team successfully demonstrated a ring-waveguide coupled structure that reduced the space required by conventional technology to less than 1/100, while also showcasing the basic operation of a gigahertz ultrasonic filter. This achievement marks a significant step towards more compact, integrated, and multifunctional ultrasonic filters for next-generation wireless communication devices.

Key Points:

• World's first gigahertz ultrasonic circuit using topological principles
• Enables stable ultrasonic wave propagation without backscattering
• Potential for miniaturization of ultrasonic filters to less than 1/100 of current size
• Applications in smartphones, IoT devices, and 5G networks
• Developed using ultrasonic topological phononic crystal material
• Demonstrates successful operation of a compact gigahertz ultrasonic filter
• Expected to lead to more efficient and high-performance wireless communication devices

https://group.ntt/en/newsrelease/2024/07/22/240722b.html

Read More

Equinix acquires 3 data centers in the Philippines

 Equinix has announced its planned entry into the Philippines through the acquisition of three data centers from Total Information Management (TIM), a prominent technology solutions provider. This strategic move, following recent expansions in Malaysia and Indonesia, aims to help businesses leverage the digital opportunities in the rapidly growing Southeast Asia region. The all-cash transaction, valued at approximately 15 times the projected EBITDA at full utilization, is expected to close in the second half of 2024, pending customary closing conditions.

The acquisition will enable Equinix to meet the digital needs of local and international businesses in the Philippines. Enterprises, cloud and IT service providers, and network service providers worldwide can use Platform Equinix® to interconnect and exchange data securely within a vibrant ecosystem of business partners and customers. TIM’s existing customers, including network and financial services companies, will also gain access to Equinix’s global ecosystem of over 10,000 companies, including more than 2,000 networks and 3,000 cloud and IT service providers. With the Philippines’ digital economy expected to grow to $35 billion by 2025, driven by a highly engaged digital population and government initiatives, this acquisition positions Equinix to support the country’s surging demand for digital infrastructure services.

Key Points:

• Acquisition includes three carrier-neutral data centers with over 1,000 cabinets and land for expansion.
• Equinix’s expansion in Southeast Asia supports digital transformation for local and global customers.
• Equinix is also expanding into Jakarta, Indonesia, and Chennai, India, and increasing its footprint in Singapore.
• Platform Equinix spans 260 data centers across 71 metros in 33 countries, including 56 data centers in 14 key metros in Asia-Pacific.
• The Philippines’ digital economy is projected to reach $35 billion by 2025, growing at a 20% CAGR.

Read More

CoreWeave brings NVIDIA Blackwell clusters to the market

CoreWeave has announced it will be among the first providers to offer large-scale NVIDIA Blackwell clusters, featuring two configurations: the NVIDIA HGX B200 and the NVIDIA GB200 NVL72. Both configurations are interconnected with NVIDIA Quantum-2 InfiniBand networking, promising significant advancements in AI, data processing, and high-performance computing.

The NVIDIA HGX B200, designed for the most demanding workloads, leverages NVIDIA Blackwell GPUs to deliver 15 times faster real-time inference on trillion-parameter models. Meanwhile, the NVIDIA GB200 NVL72 is a liquid-cooled, rack-scaled solution that integrates 36 NVIDIA Grace CPUs and 72 NVIDIA Blackwell GPUs, achieving up to 30 times faster real-time trillion-parameter LLM inference. Introduced earlier this year at NVIDIA GTC, the Blackwell architecture can link up to 72 GPUs in a single NVIDIA NVLink domain, enabling enhanced GPU-GPU communication bandwidth. This innovation supports a wide range of applications beyond AI training and inference, including quantum computing, drug discovery, and fusion energy.

Key Points:

• CoreWeave to deliver NVIDIA Blackwell clusters with configurations HGX B200 and GB200 NVL72.
• Both configurations use NVIDIA Quantum-2 InfiniBand networking for high-performance interconnectivity.
• NVIDIA HGX B200 offers 15 times faster real-time inference for trillion-parameter models.
• NVIDIA GB200 NVL72 combines 36 Grace CPUs and 72 Blackwell GPUs for 30 times faster LLM inference.
• Blackwell architecture supports extensive GPU-GPU communication for advanced applications beyond AI.

https://www.coreweave.com/

Read More

Telefónica invests in Nearby Computing for edge orchestration

Wayra, Telefónica’s corporate venture capital arm, has announced its investment in Nearby Computing, a leading startup in edge computing orchestration and automation. This investment is part of a $7 million Series A financing round aimed at accelerating Nearby Computing’s expansion and strengthening its market position.

Nearby Computing is integrated with CAMARA, an open-source project that supports Open Gateway, a global telco initiative led by the GSMA. This initiative aims to transform telecommunications networks into programmable platforms, enhancing the scalability of standardized projects. Nearby Computing’s solution streamlines the orchestration of deployments across various edge nodes, ensuring high levels of availability and low latency, which are crucial in today’s complex digital landscape.

Key Points:

• Wayra’s investment is part of a $7 million Series A funding round.
• Nearby Computing is recognized for excellence in edge computing orchestration and automation.
• The startup is integrated with CAMARA and Open Gateway, global initiatives led by the GSMA.
• Nearby Computing’s solutions ensure high availability and low latency for edge deployments.
• The investment will support the company’s expansion and market consolidation.

https://www.telefonica.com/

Read More

SECOM deploys Infinera in southeastern Colorado

SECOM, a subsidiary of Southeast Colorado Power Association (SECPA), is modernizing its middle-mile and business Ethernet access network using Infinera’s XTM Series optical transport solution. The network upgrade is set to support the bandwidth growth of large enterprises and industrial parks while providing the flexibility, reliability, and reach needed to close the digital divide in the region.

Key Points:

• The fiber network spans around 2,000 miles throughout southeastern Colorado.
• Infinera’s XTM Series will enable multi-gigabit Ethernet services and 100G/400G broadband transport.
• The upgrade enhances connectivity for homes, schools, libraries, government entities, telecoms, and businesses.
• The modernization aims to close the digital divide and maximize economic and social development opportunities.
• The XTM Series provides a temperature-hardened and low-latency packet optical middle-mile network suitable for various traffic types and demanding environments. 
• Infinera worked with its partner World Wide Technology (WWT) on the network design and deployment for SECOM.

Read More

Converge ICT picks Ribbon's Apollo for optical backbone

Converge ICT Solutions, a leading fixed broadband service provider in the Philippines, has selected Ribbon Communications’ Apollo optical networking system to optimize data transmission across its backbone network. Ribbon’s advanced 5-nanometer (nm) 140Gbaud transmission chipset will expand Converge’s fiber network capacity from 800 Gigabits per second (Gbps) to 1.2 Terabits per second (Tbps) per channel. This upgrade will offer three times the required capacity for hyperscale applications, which typically demand around 400 Gbps.

In addition to the Apollo system, Converge is leveraging Ribbon’s Muse SDN Domain Orchestration, providing the necessary toolsets for maximizing network capabilities. This includes automation, planning, node design, and real-time control, ensuring peak efficiency. “We’re thrilled to deepen our relationship with Converge ICT with this major new deployment. Coupled with our technology, our local presence and proven round-the-clock after-sales support are key factors in our ability to support Converge ICT as they ramp their network to meet increasing demands for capacity,” said Ribbon COO and EVP Sam Bucci.

Key Points:

• Converge ICT Solutions is enhancing its network with Ribbon Communications’ Apollo optical networking system.
• Ribbon’s 5nm 140Gbaud chipset will increase Converge’s network capacity from 800 Gbps to 1.2 Tbps per channel.
• The upgrade will provide three times the capacity needed for hyperscale applications.
• Converge is also using Ribbon’s Muse SDN Domain Orchestration for automation and real-time control.
• Ribbon Communications, based in Texas, offers real-time communications and IP optical networking solutions globally.
• Converge has been investing in its network infrastructure, doubling its metro backbone capacity to 800 Gbps in 2021.
• Recently, Converge signed an MOU with Super Micro Computer Inc. to establish AI-powered, green data centers in the Philippines.

Read More

Sivers Semiconductors appoints Vickram Vathulya as CEO

Sivers Semiconductors AB (Sivers), a  supplier of integrated chips and photonics modules for the most advanced communications and sensor solutions, appointed Vickram Vathulya as its new President & CEO, succeeding Anders Storm, effective August 19, 2024.

Dr. Vathulya most recently served as President of Nuvotronics, spearheading a strategic, operational, and cultural transformation for long-term growth and value creation. Prior to that, he revitalized the standard products business at Maxim Semiconductors, the largest and most profitable business in the company portfolio. His experience also includes other executive leadership roles, such as successfully growing a variety of RF and wireless businesses at Maxim Integrated and NXP Semiconductors. Vathulya has a Ph.D. in Electrical Engineering from Lehigh University in Pennsylvania and an MBA from the Berkeley Haas School of Business, University of California.

Dr. Bami Bastani, recently appointed Chairman of the Board for the Sivers Semiconductors AB, said: "Vickram's leadership qualities and industry expertise are exactly what Sivers was looking for in this CEO appointment in order to accelerate our aggressive growth strategy in the US and worldwide. His proven ability to deliver profitable growth and build high-performance teams will be invaluable as we aim to cement our position as a leader in advanced semiconductor and photonic solutions."

Read More