telecomhall

It's not a matter of if, but when and how AI will completely transform internet network traffic. Following Omdia's analysis on AI traffic, I firmly believe we are on the verge of a massive paradigm shift. AI models are primarily trained in large datacenters. However, the AI inferencing part will increasingly become a real-time activity and the need for processing at the edge will grow significantly. Deployment of AI models at the far edge will fundamentally change our approach to network traffic. Omdia expects massive growth in AI traffic: Net new AI applications, which accounted for less than 1 exabyte per month in 2023, are projected to reach nearly 150 exabytes per month by 2030. More than half of this "inferencing" traffic is expected to be handled by edge-hosted shared infrastructures. This shif will drive a strong edge computing focus, especially for applications like XR and high-quality AI video, which have substantial bandwidth requirements. Today, our networks are primarily built based on human traffic profiles. Moving forward, we must rethink and redesign our networks to accommodate the unique demands of "AI traffickers". Omdia’s analysis suggests that AI inferencing at the edge will require an average of 30Gbps capacity for far edge nodes, 80Gbps for mid-edge nodes, and a humongous 480Gbps for near-edge nodes by 2030. This transition to AI-driven traffic at the edge is not just about increasing capacity. It will require intelligent automation, orchestration capabilities, and a end to end view of network and computing resources. By 2027-2028, we must be prepared for heavy AI use cases such as XR and new media formats, which will demand a substantial different network approach. https://lnkd.in/gvPaRBm6

#handover series #1 In #LTE networks, handovers are crucial for maintaining seamless connectivity as a user equipment (UE) moves across different cell coverage areas. The two primary types of handovers are X2 and S1 handovers. These handovers involve different interfaces and processes, impacting how the network manages mobility. X2 Handover X2 Handover occurs directly between two eNodeBs (source eNodeB and target eNodeB) using the X2 interface. This handover is typically preferred when both eNodeBs are connected to the same core network and the X2 interface is available. S1 Handover S1 Handover is used when the X2 interface is unavailable, or the source and target eNodeBs are not directly connected via the X2 interface. This handover involves the S1 interface and the core network entities like the MME and the Serving Gateway (SGW). Key Parameters for Choosing the Appropriate Handover Type 1. Availability of X2 Interface: If the X2 interface is available and operational, X2 handover is preferred. If not, S1 handover is used. 2. Inter-eNodeB Connectivity: X2 handover is preferred when both eNodeBs belong to the same network operator and are within the same core network domain. S1 handover is used for inter-operator handovers or when the eNodeBs are managed by different core network domains. 3. Latency and Performance Requirements: X2 handovers are typically faster and have lower latency compared to S1 handovers, making them preferable for high-mobility scenarios. Network Topology and Configuration: 4. Network design and topology also influence the choice, with denser deployments and closely connected eNodeBs favoring X2 handovers. Advantages and Disadvantages X2 Handover Advantages: Lower Latency: Direct communication between eNodeBs reduces handover delay. Efficient Resource Management: Quick resource reallocation and less signaling overhead. Better User Experience: Faster handovers provide a seamless experience for the user. Disadvantages: Dependency on X2 Interface: Requires a functional X2 interface between eNodeBs. Limited to Intra-Network: Typically limited to eNodeBs within the same operator's network. S1 Handover Advantages: Flexibility: Can be used even when the X2 interface is not available. Inter-Operator Handover: Suitable for handovers between different network operators. Disadvantages: Higher Latency: Involves core network elements, leading to increased handover delay. Increased Signaling Overhead: More signaling messages between the core network and eNodeBs.

🛰️ Non Terrestrial Networks (NTN) take flight   👉🏽 As satellite technology continues to advance, the potential for satellite networks to supplement and enhance terrestrial networks is becoming increasingly clear. Satellite launch costs have come down significantly.   This along with cost reduction of satellite hardware has started a flurry of activity in the satellite communications space. Companies have come up with plans for ‘mega-constellations’ that have thousands of satellites.   Lower latency (via LEO constellations) offers a proposition for a competitive broadband performance in remote areas   .....plus have a capability to directly communicate with regular smartphones.   Requirements for NTN i.e. Non Terrestrial Networks were included in 3GPP Release 15 with some more additions for specific implementations in Release 17/18.   The list of broadband providers and partnerships offering IoT and D2C services only continues to grow.   Source: ABI research #satellite #5g #telecom #network #VPspeak [^474]

😂😂😩

𝗪𝗵𝗮𝘁 𝗶𝘀 𝘁𝗵𝗲 𝗗𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗕𝗲𝘁𝘄𝗲𝗲𝗻 𝗜𝗻𝗻𝗲𝗿, 𝗟𝗲𝗳𝘁, 𝗥𝗶𝗴𝗵𝘁, 𝗮𝗻𝗱 𝗙𝘂𝗹𝗹 𝗝𝗼𝗶𝗻? 🔹 (INNER) JOIN - return all rows with matching values in both tables. 🔹 LEFT (OUTER) JOIN - return all rows from the left table and those that meet the condition from the right table. 🔹 RIGHT (OUTER) JOIN - return all rows from the right table and those that meet the condition from the left table. 🔹 FULL (OUTER) JOIN - return all rows with a match in either table. Image credits: Andreas Martinson. #technology #softwareengineering #data #techworldwithmilan #sql

Today's words of wisdom.