5G
Article | September 28, 2023
Demand for data center compute continues to be strong and we believe 1Q21 would have been even stronger had it not been for the semiconductor supply shortage. We learned from vendors that the flow of server CPUs out of TSMC and Intel’s fabs was steady in 1Q21 but supply of other components necessary to build a server was tight, including power semis, BMC and PCB substrate.
Read More
Wireless, 5G
Article | May 18, 2023
For communications service providers, questions still loom over how 5G will be monetized. Many have voiced the view that the enterprise segment is where 5G will make its money, with potential use cases across a breadth of verticals including the automotive, manufacturing and utilities industries to name a few. However, if service providers focus more on delivering 5G business use cases to small and midsized enterprises (SMEs), this market segment could prove to be a fruitful source of revenue.
As 5G deployments continue to roll-out globally, CSPs are focused on deriving value from projects and partnerships with large multinational enterprises. And they are eager to do so quickly to assure payback from their $1trillion investments in the next generation network. However, are CSPs blinkered in their approach to generating ROI from 5G? Are they limiting themselves by focusing too much on a small segment of the enterprise market, that to date has limited CSPs' involvement in early 5G projects, reducing them to mere providers of connectivity?
Read More
5G
Article | May 18, 2023
Asia stands out as home to a handful of telcos busy building an international business out of selling their internally developed IT platforms. Leading the way are Jio in India, Japan’s Rakuten and Singapore operator Singtel.
Having built their own businesses, they are now selling their platforms to support new 5G business models for enterprises and other operators. In the case of Singtel, this means its 5G multi-access edge computing (MEC) services, based on Paragon, its orchestration platform for enterprise services.
Manoj Prasanna Kumar, Head of Enterprise Platforms at Singtel, who is responsible for the Paragon platform, discusses in this article the company’s enterprise service ambitions, how it’s partnering with global enterprise software vendors and the obstacles it still sees to 5G B2B service uptake.
Paragon, which falls under the telco’s DigitalInfraCo arm, aims to give enterprises “a single pane of glass that provides an end-to-end view and control of the network, the edge and the application ecosystem,” says Manoj. “It opens up the edge to the enterprise world, allowing them to deploy either their own applications or applications from Singtel's ecosystem.”
Launched last year, Paragon also lets telcos orchestrate end-to-end 5G enterprise networking services in combination with applications from software and cloud computing partners. Paragon’s application partners include Amazon Web Services, Intel, Microsoft and SAP, and the platform is available to every 5G enterprise user within the Singtel Group.
Singtel’s bet is that a growing number of enterprises will need a tightly intertwined combination of 5G connectivity and cloud computing on the edge to run specific vertical applications.
“Our strategy is to become a super aggregator of MEC,” says Manoj. “We focus on high throughput, low latency use cases, such as video analytics or streaming, mixed reality and virtual reality which pump data into the back-end applications and where the decision-making cannot afford even a few milliseconds of extra latency.”
In addition to Paragon, Singtel Group’s investments in 5G infrastructure and service delivery include a national 5G standalone (SA) network, covering more than 95% of Singapore, and international investment in data centers to support cloud computing on the network edge. Today, there are signs that its investments in 5G enterprise services are starting to bear fruit. In the second half of the 2022/23 financial year, which ended on 31 March, Singtel reported that higher demand for technology solutions and 5G services contributed to ICT revenue growth of 11%, with ICT revenues contributing 23% of Singtel Group’s overall enterprise revenue.
Singtel scored a notable win for the Enterprise 5G offering powered by Paragon platform last year when Silicon manufacturer Micron said it would deploy it and Singtel’s 5G campus network infrastructure to support its smart manufacturing operations. Micron is using Singtel’s solution to help manage and analyze its manufacturing processes for enhanced efficiency. Likewise, Singtel recently announced Hyundai as another customer for their Enterprise 5G offering powered by the Paragon platform to deliver digital twin for their electric vehicle manufacturing plant in Singapore for advanced manufacturing operations.
Nonetheless, Manoj recognizes that challenges remain when it comes to growing the 5G enterprise business. “5G and edge in Singapore have had quite a good start. But I would say we've got a long way to go,” he says.
Convincing customers
One of the biggest obstacles is generating customer demand. After all, just because enterprises are able to set 5G connectivity parameters on demand or use MEC for 5G applications at the click of a button doesn’t mean they see a reason to do so.
“Many customers don't have a lot of awareness of how edge computing can really transform their business and how a few milliseconds of latency can actually save money for them, make them more efficient, and reduce errors and so on,” says Manoj.
This reality has shaped Singtel’s sales process. “We spend quite a lot of time in raising awareness amongst customers,” he explains. “We never start with what 5G can do. Instead, we focus on understanding their challenges, their current processes, what gaps there are, and…start with applications that can help solve their problems.”
Another challenge is a lack of 5G-native devices. “This puts us in a very tough spot because when we go and connect devices to wi-fi hotspots, and then use 5G as backhaul, customers often ask ‘isn't this similar to wi fi? Why do I need 5G?’” He adds: “It will be a bit of a roadblock…for all telcos until the 5G-native device ecosystem matures.”
There is also a need for software applications that can perform optimally on 5G and the edge, and switch between network slices with different payloads. “There is a little bit of hand holding required when we bring in an ISV to qualify their application so that it can benefit from all the capabilities of 5G and the edge,” says Manoj.
And then there are the engineering challenges associated with orchestration. Paragon sets out to automate much of the orchestration and management capabilities that make it possible to request quality of service on demand for specific applications and use cases. But here again, success is dependent on close partnerships with third parties.
“Strategic partnerships with Ericsson on the network side and with Intel, Microsoft and AWS help us boost the infrastructure and the application side to stitch together the network and the infrastructure capabilities,” explains Manoj.
Choosing your vertical
Singtel is currently targeting three strategic verticals: manufacturing, public safety and urban planning. Its choice reflects the opportunities in both Singapore and the domestic markets of members of the Singtel Group.
“In Singapore, we are lucky because both enterprises and the government are very, very future-looking and invest quite a lot in adopting new technology,” says Manoj. In particular, “public sector customers are more motivated to explore something new because they carry the digital footprint of the country,” he says.
And because governments operate public safety and urban planning systems at a national level, the promises are on enough scale to spur third parties to invest in developing devices and software applications. Typical public safety use cases include video analytics, surveillance systems and robotics applications; urban planning covers systems such as traffic management.
Some of the enterprise applications Singtel sees gaining traction include immersive B2B2C content, such as delivering real-time analytics to gamers via a 360-degree video feed or mixed reality applications to train factory workers on how to troubleshoot to use complex equipment. “If they need an augmented overlay of information through the camera feeds then they need 5G and edge because a lag will make users nauseous,” explains Manoj. Other promising use cases include autonomous drones and robots.
Singtel has drawn on standard APIs, including TM Forum’s Open APIs, CAMARA APIs to build Paragon. Manoj encourages both technology standardization and collaboration with hyperscalers and software vendors to grow the enterprise market.
“Telcos should be embracing tech players as partners, seeing them as catalysts of more pull through on their services,” says Manoj. “When you partner with them, you expose your services on the hyperscale infrastructure, you naturally work with developers, which allows telcos to expand the services market.”
Read More
Network Infrastructure, Network Management
Article | July 27, 2023
Discover key network performance metrics to enhance user experience. Explore in-depth latency, throughput, jitter, packet loss, VOIP quality, and MOS score to optimize network performance analysis.
Contents
1. Importance of Network Performance Metrics for Performance Analysis
2. Critical Key Network Performance Metrics to Monitor
2.1 Latency
2.2 Throughput
2.3 Jitter
2.4 Packet Loss
2.5 VOIP Qualiy
2.6 MOS Score
3. Steps to Monitor and Measure Network Performance
4. Significance of Monitoring Metrics in Network Troubleshooting
4.1 Provides Network Visibility
4.2 Prevents Network Downtime
4.3 Observe Bandwidth Usage
5. Overcome Monitoring Challenges in Network Performance Metrics
6. Key Takeaway
1. Importance of Network Performance Metrics for Performance Analysis
Network performance involves analyzing and evaluating network statistics to determine the quality of services provided by the underlying computer network. Considering various key network metrics, it is primarily measured from the end-users’ perspective. Measuring these metrics, analyzing performance data over time, and understanding the impact on the end-user experience is essential to assess network performance.
Measuring network performance requires considering factors such as the location and timing of measurements. For instance, network performance may differ when comparing paths between cities or during periods of varying user demands throughout the day. Therefore, a comprehensive approach to monitoring network performance involves identifying these variables and identifying areas for improvement.
Network performance metrics offer valuable insights into any network infrastructure and services. These metrics provide real-time information on potential issues, outages, and errors, allowing one to allocate IT resources efficiently. Understanding end-user demands can create an adaptive network to meet future business needs. However, comprehensive monitoring requires an advanced network monitoring tool to gather, analyze, and interpret data effectively, optimizing network performance. Leveraging relevant metrics can improve network performance, help make informed decisions, enhance network reliability, and deliver a superior user experience.
2. Critical Key Network Performance Metrics to Monitor
2.1 Latency
Latency, or network delay, is a crucial performance metric in network monitoring and management. It quantifies the time required to transmit data between destinations. Factors like packet queuing and fiber optic cabling affect network latency. Consistent delays or sudden spikes in latency indicate significant network performance issues. Monitoring and minimizing latency are essential for ensuring optimal network performance. By actively tracking latency, organizations identify and address issues that may cause delays in data transmission, thereby improving overall network responsiveness and minimizing disruptions for end-users.
2.2 Throughput
Throughput metrics for network monitoring enable measurement of the data transmission rate across various network segments. Unlike bandwidth, which represents the theoretical data transfer limit, throughput reflects the successful delivery of data packets to their destination. Variations in throughput can occur across different network areas. A low throughput indicates the presence of dropped packets requiring retransmission, and highlights potential performance issues that need attention. Monitoring throughput is crucial for effective network management. By monitoring this performance metric, organizations can gain insights into the actual data transmission rate, ensuring that it aligns with expected levels.
2.3 Jitter
Jitter, a key performance metric in network monitoring, refers to the variation in delay between packets, measured as the difference between expected and actual arrival times. It results due to network congestion, routing issues, or other factors, leading to packet loss and degraded application performance. Jitter disrupts the standard sequencing of data packets and can arise due to network congestion or route changes. Monitoring jitter is crucial for identifying and addressing network stability issues and ensuring reliable data transmission. By actively monitoring this performance metric, organizations can address variations in packet delay, mitigating issues that leads to packet loss and enabling proactive troubleshooting.
2.4 Packet Loss
Packet loss, a performance management network monitoring metric, represents the number of data packets lost during transmission. It directly affects end-user services, leading to unfulfilled data requests and potential disruptions. Packet loss can arise from various factors, including software problems, network congestion, or router performance issues. Monitoring the entire process precisely to detect and address packet loss, ensures reliable data transmission and optimal network performance. Monitoring packet loss with the right network monitoring software enables timely troubleshooting and optimization of network infrastructure, ultimately enhancing overall network reliability and performance.
2.5 VOIP Quality
VoIP (Voice over Internet Protocol) quality is a crucial network performance metric. It refers to the overall performance of a VoIP system in delivering clear and reliable voice communications over the Internet, replacing traditional phone lines. Factors influencing VoIP quality include network bandwidth, latency, packet loss, jitter, and the quality of end-user devices. Monitoring VoIP quality ensures optimal system functionality and high-quality voice communications. Key performance indicators (KPIs) such as mean opinion score (MOS), jitter, latency, packet loss, and call completion rates are utilized to assess and optimize VoIP quality.
2.6 MOS Score
Mean opinion score (MOS) is a vital performance metric in network monitoring, rating the perceived quality of a voice call on a scale of 1 to 5. It is a standardized measurement developed by the ITU, an international agency focused on enhancing communication networks. Initially designed for traditional voice calls, the MOS has been adapted to evaluate Voice over IP (VoIP) calls. The MOS score considers various factors, including the specific codec employed for the VoIP call, providing a comprehensive assessment of voice calls quality in network monitoring.
3. Steps to Monitor and Measure Network Performance
Step 1: Deploy a Software for Network Monitoring
To effectively measure network performance, deploying dedicated network monitoring software is crucial. While temporary tools like traceroutes and pings can provide insights into ongoing problems, they are insufficient for troubleshooting intermittent network issues. Relying on periodic tools for intermittent issues is reliant on chance, as it may only detect problems when they occur during tool usage. By implementing comprehensive network monitoring software, one can proactively monitor and analyze network metrics, historical data, and performance, allowing for timely detection and resolution of both ongoing and intermittent network issues.
Step 2: Distribute Monitoring Agents
For comprehensive network performance measurement, businesses must distribute monitoring agents strategically across key network locations. These specialized software agents continuously monitor network performance using synthetic traffic, simulating and assessing the end-user perspective.
By distributing Monitoring Agents, organizations can:
• Measure key network metrics, including jitter, packet loss, and throughput.
• Identify and troubleshoot intermittent network issues that are challenging to pinpoint.
• Receive alerts regarding any performance degradation, ensuring a timely response.
• Collect valuable data for in-depth troubleshooting and analysis, facilitating proactive network management and optimization.
Step 3: Measure Network Metrics
After deploying the monitoring agents, they continuously exchange synthetic User Datagram Protocol (UDP) traffic, forming a network monitoring session. During this session, the agents measure network performance by evaluating key metrics and conducting network traffic analysis. The metrics used in the analysis include specific parameters, and the results of these measurements are presented in a network response time graph, providing a visual representation of the network's performance characteristics. Monitoring and analyzing these metrics enable organizations to gain valuable insights into network performance, facilitating informed decision-making and convenient network performance troubleshooting.
4. Significance of Monitoring Metrics in Network Troubleshooting
4.1 Provide Network Visibility
Monitoring metrics plays a vital role in network troubleshooting by offering network visibility. They enable the identification of performance bottlenecks, configuration problems, and security vulnerabilities that detrimentally affects network performance. These issues can be addressed through targeted troubleshooting efforts, resulting in improved network performance and enhanced end-user experience. Organizations identify and resolve network issues by monitoring metrics, ensuring optimal network functionality and overall business productivity.
4.2 Prevent Network Downtime
Effective monitoring metrics are instrumental in preventing network downtime, a costly concern for businesses. Swift identification and resolution of network issues through proactive network performance troubleshooting help minimize downtime, ensuring uninterrupted business operations. By promptly addressing potential problems, network troubleshooting safeguards against lost productivity, revenue, and customer dissatisfaction. Maintaining a proactive approach to monitoring and resolving network issues to enhance network reliability and business continuity.
4.3 Observe Bandwidth Usage
Monitoring metrics are essential in network troubleshooting as they enable the observation of bandwidth usage. This allows organizations to detect abnormal or excessive utilization, pinpoint key performance issues and ensure optimal resource allocation. It allows for identifying critical bandwidth-hogging applications or network intrusions, helping experts take immediate action to mitigate risks, safeguard data, and protect the overall network integrity. Additionally, experts can optimize network performance and ensure a seamless user experience for organizations relying on efficient network infrastructure.
5. Overcome Monitoring Challenges in Network Performance Metrics
Enterprises seeking to ensure optimal network performance and improve overall business operations must overcome network monitoring obstacles. Effectively monitoring, tracking, and improving network performance requires a strategic combination of skilled personnel, advanced technologies, and well-defined strategies. Failing to address these requirements results in various challenges that hinder the ability to enhance network performance effectively.
The challenges that businesses often encounter include managing scalability, handling massive data volumes, achieving real-time monitoring, dealing with multi-vendor environments, addressing network security and privacy concerns, and adapting to evolving network demands. Each obstacle presents unique complexities that require tailored approaches and expert insights.
To overcome these challenges, enterprises must invest in comprehensive monitoring tools capable of handling the scalability demands of growing networks. These tools should provide real-time network visibility, robust analytics capabilities, and intelligent data filtering mechanisms to extract meaningful insights from vast network data. Establishing clear monitoring objectives aligned with business goals and defining key performance indicators (KPIs) are essential in effectively addressing network performance challenges.
6. Key Takeaway
Monitoring network performance metrics is crucial for assessing the quality of services a computer network provides from an end-user perspective. It involves continuously tracking and analyzing key metrics such as latency, throughput, jitter, packet loss, VOIP quality, and MOS score. Organizations can actively monitor and assess performance, proactively identify intermittent issues, and collect valuable data for in-depth analysis by implementing dedicated network monitoring software and strategically deploying monitoring agents across the network. In addition, it is imperative to emphasize the significance of monitoring metrics in mitigating the potential financial impact of network downtime, enhancing the utilization of available bandwidth resources, and efficiently tackling the complexities inherent in scaling operations, real-time monitoring, diverse vendor ecosystems, security concerns, and the ever-evolving requirements of modern networks.
Read More