# Routing Published articles for Routing. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## MCP went stateless: What changed in the 2026-07-28 spec DevFeed: [MCP went stateless: What changed in the 2026-07-28 spec]() Original publisher: [Read original article]() Author: WorkOS Published: 2026-09-16T00:00:00Z Content type: article Language: en Sources: [WorkOS Blog]() Topics: [Model Context Protocol](), [migration](), [HTTP](), [servers](), [client]() Tags: [api](), [client](), [http](), [load-balancer](), [mcp](), [replacement](), [request](), [routing](), [server](), [servers](), [sessions](), [stateless](), [stream]() ### AI overview The 2026-07-28 MCP specification removes protocol-level sessions, the initialize handshake, and stream resumability, making the protocol behave more like an ordinary HTTP API. Most migrations are mechanical, but lost stream requests can cause reliability regressions or duplicated side effects when retried, so side-effecting tools need application-level idempotency keys. The elicitation change also requires architectural rework. ### Source excerpt Sessions, the initialize handshake, and stream resumability are all gone. Here is what replaced them, and what breaks if you ignore it. ## Dense vs. MoE Models: Active Parameters, Throughput, and When to Choose Each DevFeed: [Dense vs. MoE Models: Active Parameters, Throughput, and When to Choose Each]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-09-15T17:00:11Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Mixture of Experts (MoE)](), [Nemotron](), [Routing (disambiguation)]() Tags: [agentic-ai-generative-ai](), [ai-inference](), [architecture](), [cost](), [data-center-cloud](), [deployment](), [llms](), [memory](), [mixture-of-experts-moe](), [models](), [moe](), [nemo](), [nemotron](), [performance](), [router](), [routing](), [tokens]() ### AI overview This article explains how dense and Mixture-of-Experts models activate parameters, compares their effects on throughput, memory cost, and serving complexity, and discusses when each architecture fits different deployment constraints. It uses Nemotron 3.5 Lightning as an example of an MoE model. ### Source excerpt How can a 30B-parameter model activate only 3B parameters per token, and still use the capacity of the larger model? Nemotron 3.5 Lightning illustrates the... ## Replica-aware routing public beta DevFeed: [Replica-aware routing public beta]() Original publisher: [Read original article]() Author: Amy Chen; Jan Mensch Published: 2026-09-15T13:15:00Z Content type: release Language: en Sources: [ClickHouse Blog]() Topics: [clickhouse](), [Routing (disambiguation)](), [Replication](), [Caching]() Tags: [cache](), [clickhouse](), [consistency](), [replication](), [routing]() ### AI overview ClickHouse introduces replica-aware routing in Public Beta for Enterprise customers. The feature routes requests to the same replica, allowing continued access to temporary tables and named sessions and supporting read-after-write consistency over HTTP or the native protocol. ### Source excerpt Temporary tables and named sessions live on a single ClickHouse replica, so a follow-up query routed elsewhere can't see them. Replica-aware routing pins your requests to the same replica over HTTP or the native protocol -- and here's how we built it. ## Accelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine DevFeed: [Accelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine]() Original publisher: [Read original article]() Author: Tanya Lenz Published: 2026-09-14T16:39:15Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Mixture of Experts (MoE)](), [Artificial Intelligence](), [GPU](), [Large Language Model](), [Nvidia](), [Routing (disambiguation)](), [GB200](), [Network](), [Python](), [deepseek](), [mixtral](), [qwen]() Tags: [agentic-ai-generative-ai](), [ai](), [communication](), [deepseek](), [developer-tools-techniques](), [gb200](), [gpu](), [mixtral](), [mixture-of-experts-moe](), [mlops](), [moe](), [networks](), [nvfp4](), [nvidia](), [performance](), [python](), [qwen](), [router](), [routing](), [tensors](), [token](), [tokens](), [training-ai-models]() ### AI overview This article explains how NVIDIA Transformer Engine and JAX kernel optimizations accelerate dropless Mixture of Experts (MoE) training. It describes bottlenecks from token routing, expert dispatch and gathering, all-to-all communication, and ragged expert matrix operations. In DeepSeek-V3 training on NVIDIA GB200, the optimized approach increased performance from 103 to 1,068 TFLOPS per GPU, a 10.4x improvement. ### Source excerpt Mixture of experts (MoE) has become one of the defining architectural trends in large-scale AI model training. DeepSeek, Qwen, and Mixtral are examples of MoE... ## Chinese AI models dominate OpenRouter's US token consumption. It can now guarantee that traffic stays entirely in the US. DevFeed: [Chinese AI models dominate OpenRouter's US token consumption. It can now guarantee that traffic stays entirely in the US.]() Original publisher: [Read original article]() Author: Paul Sawers Published: 2026-09-14T13:59:33Z Content type: news Language: en Sources: [The New Stack]() Topics: [AI Models](), [Routing (disambiguation)](), [Security](), [data](), [Availability](), [hugging face](), [Nvidia](), [Nemotron](), [OpenAI]() Tags: [ai](), [ai-infrastructure](), [ai-models](), [availability](), [data](), [hugging-face](), [large-language-models](), [models](), [nemotron](), [nvidia](), [openai](), [routing](), [security]() ### AI overview OpenRouter has launched US in-region routing for business and enterprise customers. Requests sent through its US endpoint are decrypted, processed, and served entirely inside the United States, or rejected if that cannot be guaranteed. The feature addresses concerns about data location when businesses use open-weight models, including models developed in China. ### Source excerpt Everyone knows the open-weight model pitch by now: companies can download the weights, customize them, run them on infrastructure of The post Chinese AI models dominate OpenRouter's US token consumption. It can now guarantee that traffic stays entirely in the US. appeared first on The New Stack. ## Solana: Building, Proving and Earning Trust in Public DevFeed: [Solana: Building, Proving and Earning Trust in Public]() Original publisher: [Read original article]() Author: Jacob Creech Published: 2026-09-14T11:00:00Z Content type: article Language: en Sources: [Solana News Feed]() Topics: [Network](), [Critical Infrastructure](), [incident](), [networking](), [Transactions](), [Amazon Web Services]() Tags: [2026](), [aws](), [blockchain](), [blockchain-technology](), [building](), [critical-infrastructure](), [crypto](), [crypto-news](), [cryptocurrency](), [defi](), [ecosystem](), [incident](), [infrastructure](), [network](), [nfts](), [outages](), [podcasts](), [routing](), [solana](), [solana-ecosystem](), [technology](), [transactions](), [uptime](), [web3]() ### AI overview Solana describes how repeated public testing, transparent incident reporting, and corrective engineering have strengthened trust in the network. It highlights a 2026 routing failure that took nearly 29% of network stake offline while blocks and transactions continued, followed by recovery of the infrastructure provider in just over 30 minutes. The article also cites validator capacity improvements, stake-weighted quality of service, and a redesigned transaction scheduler as examples of resilience work. ### Source excerpt Solana has maintained 100% uptime since February 2024, including when a routing failure took nearly 29% of network stake offline. ## M5Stack Module13.2 LoRa-1262 expansion board integrates 14x10 mm Stamp LoRa-1262 module DevFeed: [M5Stack Module13.2 LoRa-1262 expansion board integrates 14x10 mm Stamp LoRa-1262 module]() Original publisher: [Read original article]() Author: Debashis Das Published: 2026-09-13T05:02:54Z Content type: news Language: en Sources: [CNX Software - Embedded Systems News]() Topics: [lora](), [systems]() Tags: [connectivity](), [development-board](), [embedded-systems](), [hardware](), [i2c](), [io](), [iot](), [lora](), [lorawan](), [low-power](), [lpwan](), [m5stack](), [module](), [news](), [power-management](), [routing](), [series](), [smart-agriculture](), [smart-city](), [spi](), [wireless]() ### AI overview M5Stack has launched the Stamp LoRa-1262 SMD module and the stackable Module13.2 LoRa-1262 expansion board for M5Stack Core controllers. Based on the Semtech SX1262, the module supports several modulation modes across 868-923 MHz, with up to +22 dBm transmit power and -147 dBm receive sensitivity. Module13.2 adds an RP-SMA antenna interface, M5IOE1 I/O expansion, configurable pin routing, and selectable I2C addresses for multi-module stacking. ### Source excerpt After launching the SX1262-based M5Stamp C6LoRa tiny module, M5Stack has launched two SX1262-based LoRa add-ons for the 868-923 MHz band: the tiny Stamp LoRa-1262 SMD module for custom PCBs, and the stackable Module13.2 LoRa-1262 module for the M5Stack Core series. The Stamp LoRa-1262 supports LoRa, FSK, GFSK, MSK, GMSK, and OOK modulation, with up to +22 dBm transmit power and -147 dBm receive sensitivity. The Module13.2 adds an RP-SMA antenna interface, an M5IOE1 IO expansion chip for reset and power management, and DIP switches for flexible pin routing and configurable I2C addresses, enabling multi-module stacking. Stamp LoRa-1262 SMD module M5Stack Stamp LoRa-1262 specifications: LoRa Transceiver - Semtech SX1262 Wireless Connectivity (LoRa) Frequency Band - 868 to 923 MHz Modulation Modes - LoRa, FSK, GFSK, MSK, GMSK, OOK Bitrate - Up to 300 kbps (programmable) Transmit power (Tx) - Up to +22 dBm Receive sensitivity (Rx) - Down to -147 dBm [...] The post M5Stack Module13.2 LoRa-1262 expansion board integrates 14x10 mm Stamp LoRa-1262 module appeared first on CNX Software - Embedded Systems News. ## Scaling AI Agent Infrastructure with the MCP Stateless updates DevFeed: [Scaling AI Agent Infrastructure with the MCP Stateless updates]() Original publisher: [Read original article]() Author: Kurtis Van Gent; Alan Blount Published: 2026-09-12T11:04:33.891311Z Content type: release Language: en Sources: [Google Developers Blog]() Topics: [networking](), [SDKs]() Tags: [agentic](), [ai-agent](), [c-sharp](), [go](), [http](), [infrastructure](), [mcp](), [python](), [routing](), [sdks](), [serverless](), [typescript](), [updates]() ### AI overview The MCP release candidate replaces transport-level session management with a stateless core for HTTP load balancing, cloud-native scaling, and serverless deployments. It also introduces routing headers, caching controls, MRTR, and beta SDKs for Python, TypeScript, Go, and C#. ### Source excerpt The 2026-07-28 Model Context Protocol (MCP) specification replaces legacy stateful constraints with a fully stateless core, enabling cloud-native horizontal scaling, serverless deployments, and standard round-robin load balancing. This architectural shift introduces standardized HTTP headers for efficient routing without deep packet inspection, caching controls, and Multi Round-Trip Requests (MRTR) to handle interactive and long-running tasks without blocking connections. Developers can immediately begin migrating their agentic applications to this highly scalable infrastructure using the newly available beta SDKs for Python, TypeScript, Go, and C#. ## 4 engineering patterns behind the strongest AI Agents Challenge submissions DevFeed: [4 engineering patterns behind the strongest AI Agents Challenge submissions]() Original publisher: [Read original article]() Author: Sergio Villani Published: 2026-09-12T11:04:33.891311Z Content type: article Language: en Sources: [Google Developers Blog]() Topics: [AI Bots](), [MSP MCP]() Tags: [agents](), [ai-agents](), [cost](), [engineering](), [event](), [inference](), [latency](), [mcp](), [mcp-server](), [routing](), [validation]() ### AI overview An article on engineering patterns found in leading AI Agents Challenge submissions. It highlights bidirectional MCP, asynchronous event buses, unified validation for model fallbacks, and tiered routing to make agentic workflows more resilient, faster, and less costly. ### Source excerpt The recent Google for Startups AI Agents Challenge revealed that the most successful multi-agent systems rely on foundational software engineering patterns rather than just raw model power. Winning architectures consistently implemented bidirectional MCP for seamless inter-agent communication, async event buses for parallel execution, strict unified validation for model fallbacks, and tiered routing to minimize expensive inference calls. By prioritizing these structural practices over simple linear prompt chains, developers can build more resilient, low-latency, and cost-effective agentic workflows. ## Scaling real-time AI agents with session-aware load balancing DevFeed: [Scaling real-time AI agents with session-aware load balancing]() Original publisher: [Read original article]() Author: Simerus Mahesh Published: 2026-09-12T11:04:33.891311Z Content type: article Language: en Sources: [Google Developers Blog]() Topics: [real-time](), [Artificial Intelligence](), [Streaming](), [Server](), [gRPC](), [WebSocket](), [cpu](), [Concurrent Programming]() Tags: [ai-agents](), [concurrent](), [cpu](), [real-time](), [routing](), [server](), [streaming]() ### AI overview This article explains why real-time AI agents require session-aware load balancing. Long-lived, stateful bidirectional streams make request rates and CPU utilization insufficient measures of backend capacity. The proposed approach tracks active sessions at the application level and combines session counts with CPU metrics to distribute traffic and avoid bottlenecks. ### Source excerpt Real-time AI agents break traditional request-response load balancing paradigms because they rely on long-lived, stateful bidirectional streams that obscure true server capacity. To solve this, developers must implement application-level session tracking directly within the runtime to accurately measure the committed concurrent workload of active conversations. By feeding these precise session counts alongside standard CPU utilization metrics into a hybrid routing algorithm, infrastructure can effectively distribute stateful AI traffic and prevent individual backend bottlenecks. ## Your built-in router VPN might be more trouble than it's worth DevFeed: [Your built-in router VPN might be more trouble than it's worth]() Original publisher: [Read original article]() Author: Andrew Cunningham Published: 2026-09-11T16:00:00Z Content type: opinion Language: en Sources: [Blog on Tailscale]() Topics: [networking](), [Networks](), [Server](), [Hardware]() Tags: [nas](), [networks](), [performance](), [routing](), [security](), [server]() ### AI overview The article argues that built-in VPN servers on routers and NAS devices can be unreliable and difficult to use, especially on restrictive networks. It presents Tailscale as a better option for addressing those connection limitations. ### Source excerpt Built-in VPNs are easy to start--then harder to live with. ## Reduce LLM latency with prefix-aware routing on Amazon SageMaker Inference DevFeed: [Reduce LLM latency with prefix-aware routing on Amazon SageMaker Inference]() Original publisher: [Read original article]() Author: Kareem Syed-Mohammed Published: 2026-09-10T21:58:09Z Content type: release Language: en Sources: [Artificial Intelligence]() Topics: [Large Language Model](), [Low-Latency Inference](), [Caching](), [Inference](), [Inference Performance]() Tags: [advanced-300](), [amazon-sagemaker](), [amazon-sagemaker-ai](), [announcements](), [benchmarks](), [caching](), [inference](), [latency](), [llama](), [llm](), [routing](), [tensorrt-llm](), [vllm]() ### AI overview Amazon SageMaker Inference introduces prefix-aware routing for LLM requests. By consistently sending requests with matching prompt prefixes to the same instance, it improves reuse of cached KV computations and can reduce time to first token. ### Source excerpt Amazon SageMaker Inference now offers prefix-aware routing, a routing strategy that sends requests sharing the same prompt prefix to the same instance so the KV cache stays warm. In benchmarks on Llama 3.1 70B, it reduced P50 time-to-first-token by up to 77% and raised KV cache hit rates from about 25% to over 80%. ## Agent Gateway: The Next Evolution of the API Gateway DevFeed: [Agent Gateway: The Next Evolution of the API Gateway]() Original publisher: [Read original article]() Author: Gbadebo Bello Published: 2026-09-10T16:00:00Z Content type: article Language: en Sources: [Postman Blog]() Topics: [API](), [AI Agent](), [Architecture & Design](), [Artificial Intelligence](), [Routing (disambiguation)](), [Authentication](), [Authorization](), [observability]() Tags: [agent-gateway](), [ai](), [ai-agents](), [api](), [api-101](), [api-gateway](), [architecture](), [authorization](), [editor-s-picks](), [fabric-gateway](), [general](), [llm-gateway](), [mcp](), [mcp-gateway](), [observability](), [routing]() ### AI overview The article explains the emergence of the Agent Gateway as an architectural component for AI agents. It contrasts traditional API Gateways, which authenticate, govern, route, and deliver requests from clients that already know which API to call, with the needs of AI agents, which receive goals and may require broader governance across agentic workflows. ### Source excerpt API gateways are evolving for AI agents. Learn how the agent gateway governs identity, tools, memory, and policy across agentic workflows. Get early access. The post Agent Gateway: The Next Evolution of the API Gateway appeared first on Postman Blog. ## Introducing Neki DevFeed: [Introducing Neki]() Original publisher: [Read original article]() Author: Nick Van Wiggeren Published: 2026-09-10T12:00:00Z Content type: release Language: en Sources: [Blog -- PlanetScale]() Topics: [Database](), [MySQL]() Tags: [database](), [mysql](), [neki](), [platform](), [postgres](), [product](), [routing](), [scale](), [workflow]() ### AI overview PlanetScale announces Neki in platform preview, a sharded Postgres offering designed to scale a database across multiple machines while retaining Postgres compatibility. ### Source excerpt Neki, sharded Postgres by PlanetScale, is now available in platform preview. ## How we cut CDN metadata lookup latency by 91% DevFeed: [How we cut CDN metadata lookup latency by 91%]() Original publisher: [Read original article]() Author: Luba Kravchenko Published: 2026-09-10T00:00:00Z Content type: article Language: en Sources: [Vercel News]() Topics: [Latency](), [Vercel](), [Caching](), [Routing (disambiguation)](), [API](), [React Native]() Tags: [api](), [cache](), [cdn](), [deployment](), [latency](), [routing](), [vercel]() ### AI overview Vercel describes how it reduced P99 CDN metadata lookup latency by 91%. The team replaced per-path metadata fetching with bounded shards that group metadata for many paths, improving cache reuse across routing lookups while controlling data-fetching costs. The article also explains how framework routing rules, Build Output API outputs, Bloom filters, and exact metadata lookups work together to select the correct response. ### Source excerpt Every request to Vercel passes through our CDN, which executes on average over 80 million routing instructions per second. Part of that work is looking up metadata to determine which paths exist and how to serve them. When that metadata isn't cached, the CDN has to fetch it before it can serve the response. We used to fetch and cache metadata one path at a time, retrieving only what each lookup needed. That seemed efficient at the time, but large deployments can contain hundreds of thousands of paths, each with its own cache entry. Every new deployment introduced fresh metadata, making cache misses a recurring cost for large sites that deployed frequently. Fetching more metadata at once made these lookups faster. By grouping paths together, each fetch could populate the cache for many subsequent lookups. But fetching too much introduced its own costs. Through production experiments, we found a balance that cut P99 metadata lookup latency by 91% and made deployments faster along the way. How the CDN finds the right route The path in a request doesn't always match the path of the content or function that serves it. Framework routing rules connect the two. A request for /blog/hello-world, for example, might resolve to the dynamic route /blog/[slug]. A request for that page's React Server Component payload might resolve to /blog/[slug].rsc. As the CDN applies these rules, it may need to check several target paths to find the right response. The framework describes those routes during a build. With framework-defined infrastructure, application code declares which outputs are static, which need Functions, and which responses can be cached or regenerated. Frameworks translate that intent into Build Output API outputs, and Vercel generates the routing metadata the CDN reads. At request time, the CDN needs to determine which target paths exist and retrieve their metadata. The Bloom filters we added to global routing rule out paths that definitely don't exist. Remaining paths ## The lifecycle of a sharded Postgres query DevFeed: [The lifecycle of a sharded Postgres query]() Original publisher: [Read original article]() Author: PlanetScale Published: 2026-09-10T00:00:00Z Content type: tutorial Language: en Sources: [Blog -- PlanetScale]() Topics: [Database](), [networking]() Tags: [architecture](), [auth](), [database](), [engineering](), [neki](), [postgres](), [routing](), [scale](), [server](), [sql]() ### AI overview An overview of how a SQL query moves through a sharded Postgres database, from authentication and routing to execution across shards. ### Source excerpt Follow a SQL query through the router, across four Postgres shards, and back. ## Architecting SASE solutions using AWS Local Zones DevFeed: [Architecting SASE solutions using AWS Local Zones]() Original publisher: [Read original article]() Author: Lakshmi VP Published: 2026-09-09T20:36:09Z Content type: tutorial Language: en Sources: [AWS Compute Blog]() Topics: [networking](), [Network Services]() Tags: [advanced-300](), [architecture](), [aws](), [aws-local-zones](), [aws-marketplace](), [best-practices](), [edge](), [latency](), [routing](), [sase](), [security](), [zero-trust]() ### AI overview The article explains how AWS Local Zones can support SASE deployments by placing virtual security appliances closer to distributed users to reduce latency while maintaining security controls. ### Source excerpt Organizations with geographically distributed workforces face a trade-off between security and low-latency access. This post explores how to use AWS Local Zones and Secure Access Service Edge (SASE) solutions to deploy virtual security appliances closer to end users, covering key design principles, capacity planning, and traffic routing. ## How Smart Model Routing Can Cut LLM Costs by 10X DevFeed: [How Smart Model Routing Can Cut LLM Costs by 10X]() Original publisher: [Read original article]() Author: ByteByteGo Published: 2026-09-09T15:30:26Z Content type: article Language: en Sources: [ByteByteGo]() Topics: [Model Routing](), [Large Language Model](), [API]() Tags: [cost](), [llm](), [model-routing](), [models](), [routing](), [routing-system](), [tokens]() ### AI overview This article explains how smart model routing can reduce LLM application costs by sending simple requests to less expensive models and difficult requests to more capable models. It notes that savings and response quality depend on request types, model price differences, and routing performance, and discusses router, cascading, semantic, and learned routing approaches. ### Source excerpt Cost reduction isn't a given. It also depends on the types of requests the application receives, the price difference between models, and how well the routing system performs. In this article, we are going to look at various aspects ## Tailscale Kubernetes Operator 1.102: In-cluster Peer Relays, better IPv6, and optimized certificates DevFeed: [Tailscale Kubernetes Operator 1.102: In-cluster Peer Relays, better IPv6, and optimized certificates]() Original publisher: [Read original article]() Author: Mike Stefaniak Published: 2026-09-09T14:00:00Z Content type: release Language: en Sources: [Blog on Tailscale]() Topics: [Kubernetes](), [networking](), [VPC]() Tags: [kubernetes](), [networks](), [release](), [routing](), [vpc]() ### AI overview Tailscale Kubernetes Operator 1.102 adds in-cluster peer relays, improves IPv6 connectivity support, and reduces certificate-related deployment and connectivity issues. ### Source excerpt Faster cross-cluster connections, fewer IPv6 dead ends, less certificate churn. ## APNIC and NIXI partner to strengthen routing security and technical capacity in India DevFeed: [APNIC and NIXI partner to strengthen routing security and technical capacity in India]() Original publisher: [Read original article]() Author: Dan Fidler Published: 2026-09-09T09:01:06Z Content type: news Language: en Sources: [APNIC Blog]() Topics: [Routing Security](), [Security](), [Internet](), [Resilience](), [networking]() Tags: [apnic-62](), [collaboration](), [development](), [india](), [infrastructure](), [internet](), [ipv6](), [knowledge-sharing](), [nixi](), [partnership](), [resilience](), [routing](), [routing-security](), [rpki](), [security](), [skills](), [support]() ### AI overview APNIC and NIXI have signed an MoU to strengthen routing security and technical capacity in India. The partnership will expand IPv6 and RPKI deployment, pilot an RPKI repository mirror, promote routing security practices, and support technical training and knowledge sharing. ### Source excerpt The partnership will expand IPv6 and RPKI deployment including piloting an RPKI Repository Mirror to support a more secure and resilient Internet ecosystem. ## Deployment step now 10% faster DevFeed: [Deployment step now 10% faster]() Original publisher: [Read original article]() Author: Javi Velasco Published: 2026-09-08T22:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [CI/CD]() Tags: [manifest](), [routing]() ### AI overview Vercel made its deployment step about 10% faster by combining per-function routing metadata into one manifest upload. The improvement applies automatically to all builds. ### Source excerpt The deployment step is now about 10% faster, saving one second on average. Large applications can save up to 12 seconds. Previously, Vercel uploaded a separate routing metadata file for each function path. That metadata is now combined into a single manifest and uploaded once. This improvement applies automatically to all builds. No changes are required. Learn more about builds in the documentation. Read more ## Vercel Sandbox routing is now 18x faster globally DevFeed: [Vercel Sandbox routing is now 18x faster globally]() Original publisher: [Read original article]() Author: Andy Waller Published: 2026-09-08T20:00:00Z Content type: release Language: en Sources: [Vercel News]() Topics: [networking](), [SDKs]() Tags: [latency](), [routing](), [sandbox](), [vercel]() ### AI overview Vercel Sandbox public-domain routing now resolves domains from the nearest regional replica, reducing median lookup latency from 62 ms to 3.4 ms. The change applies automatically and has no pricing changes. ### Source excerpt Requests to Vercel Sandbox public domains are now routed 18x faster. Domains created with the sandbox.domain() SDK call are now resolved from the nearest regional replica, instead of a single centralized store. Incoming requests reach the process running in the sandbox with less latency. Median domain lookup latency dropped from 62ms to 3.4ms (18x faster). The improvement is largest in regions farthest from the previous store: lookups are now up to 112x faster in Sydney (syd1) and up to 146x faster in Cape Town (cpt1), at p99. This applies automatically to every request to a sandbox domain. There are no pricing changes. Learn more about Vercel Sandbox in our documentation. Read more ## LLM Routing Can Cost More Than Not Routing DevFeed: [LLM Routing Can Cost More Than Not Routing]() Original publisher: [Read original article]() Author: Avi Chawla Published: 2026-09-07T20:30:58Z Content type: article Language: en Sources: [Daily Dose of Data Science]() Topics: [Model Routing](), [Large Language Model](), [Frontier Model](), [coding]() Tags: [agent](), [cost](), [llm](), [model](), [models](), [routing]() ### AI overview The article examines why routing requests among LLMs can cost more than using one model throughout an agent session. Although cheaper models may handle simpler tasks, routing adds overhead and can be counterproductive inside agent loops. ### Source excerpt ...covered with a production-grade router for LLM apps. ## Indian backbones need AS\_PATH filters DevFeed: [Indian backbones need AS\_PATH filters]() Original publisher: [Read original article]() Published: 2026-09-07T20:11:07Z Content type: opinion Language: en Sources: [Personal blog of Anurag Bhatia]() Topics: [BGP](), [Routing (disambiguation)](), [networking](), [Security]() Tags: [airtel](), [as4755](), [as55836](), [as9498](), [bgp](), [india](), [jio](), [network](), [rcom](), [routing](), [tata-comm]() ### AI overview The article argues that route leaks are a longstanding security problem for India's routing infrastructure. It explains limitations of IRR AS-SETs, discusses ASPA and AS_PATH filters or peer lock, and questions whether major Indian backbones can consistently apply those filters because of indirect routing and smaller downstream networks. ### Source excerpt The security of India's routing infrastructure is something that isn't discussed as much as cybersecurity, software vulnerabilities, etc. Part of the blame for that has to go to our industry, which is relatively closed when it comes to sharing incidents and also lacks good documentation on the deployment side of things. Some of these things came to light after the RCom (AS18101) BGP hijack of Telegram prefixes in June 2026. One silent problem for a long time has been route leaks. BGP route leak Route leaks happen when a network ends up announcing a route to a BGP adjacency where it is not supposed to announce that route. Take, for example, a network "leaking" routes learnt from a peer to transit, or vice versa. One can filter downstreams which are small, but it's very hard to filter large downstream networks which have further downstreams, or if one is far removed from that ASN. IRR AS-SETs exist for this reason, but they have not worked well due to tooling challenges. Furthermore, IRR (Internet Routing Registry) by design is a public register where one "publishes intent" before actually doing that in BGP, and then anyone can match the intent with the state in BGP. New tech ASPA will help address this issue, but it's new, vendor support is still in the development phase and it will take its own adoption time. AS_PATH filters / Peer lock This is older tech which can play a very effective role in controlling leaks. If network A decides to peer with network B, they both can agree to announce all routes to each other over the peering & agree to reject each other's ASN from all other BGP sessions. It's a common practice between large transit-free Tier-1 networks as well as major backbones and is often referred to by the fancy name of "peer lock". In India, I doubt backbones like Airtel, Jio, and Tata Comm can have AS_PATH filters rejecting each other from everywhere except direct sessions because there is some indirect routing visible all the time. It could be because of [Next page]()