Loading…
Topic
378 posts about Scalability. Every summary links to the original.
See the idea
Consolidating agent memory, retrieval, and governance onto MongoDB Atlas removes sync and vendor layers, but ties scaling and control to one platform. This trade-off view shows the payoff, the cost, and the condition that decides it.
Autoscaling trades responsiveness, stability, and cost: this explainer shows how signal choice, scaling direction, and capacity limits shape the right control loop for a workload.
Apache Pinot is a distributed real-time OLAP datastore for low-latency analytics. This architecture view shows how streams, processing, partitioning, indexed storage, and replica-aware serving fit together.
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
Newest first
Dropbox manages growing infrastructure demands from AI and standard workloads by maximizing the efficiency of its existing data centers rather than just building more capacity. Through a system-level approach spanning software, hardware, and physical facility design, the company dynamically manages power and balances workloads. These optimizations, combined with higher storage density and data-driven hardware lifecycle management, have significantly reduced energy use per unit of storage.
UMAP workflows on tens to hundreds of millions of vectors are constrained by the cost of constructing an all-neighbors k-nearest-neighbor graph, especially when repeated runs support exploration and parameter tuning. The NVIDIA cuML and cuVS 25.06 feature distributes this graph-construction stage across multiple GPUs while retaining the out-of-core design that partitions data into balanced clusters, overlaps vectors across nearby clusters, computes local graphs, and merges them into a global graph. knn_n_clusters and knn_overlap_factor expose the memory, runtime, and graph-quality trade-off: more clusters reduce per-GPU data, while more overlap can improve nearest-neighbor recall at higher computational and memory cost. On eight NVIDIA H100 GPUs, the reported evaluation found up to 74x end-to-end speedup over a projected CPU runtime for the 106M-vector MIRACL dataset, completing it in 8 minutes. The post states that performance continues to improve from one to eight GPUs while preserving comparable embedding quality across configurations.
Sony LIV’s legacy analytics environment combined batch pipelines, Elasticsearch, BigQuery, and other telemetry stores, creating slow, costly, fragmented analysis as streaming activity reached billions of events per day. During live sports, the platform needed sub-second queries, high ingestion throughput, cost-efficient long-term storage, and horizontal scaling for sudden traffic spikes. Sony LIV consolidated these workloads on ClickHouse Cloud: client, streaming, CDN, and advertising systems send events through Amazon Kinesis and ClickPipes into a ClickHouse cluster serving dashboards, product analytics, QoE, and SRE workflows. The source reports that queries previously taking tens of seconds to minutes now run in sub-second to low-single-digit seconds, improving operational responsiveness, engineering productivity, and the handling of major live-event traffic.
Retailers are moving from isolated generative-AI pilots to enterprise-wide adoption, where each store, headquarters, and digital workflow needs different data, permissions, models, and cost controls. The post argues that useful AI depends on governed business context, while fragmented tools can produce separate provider contracts, logs, permission models, cost centers, and answers to the same question. It proposes an open control plane that centralizes governance and model access without forcing employees into one application or vendor. The described Databricks design combines Unity Catalog for data and business context, Unity AI Gateway for routing and oversight, Foundation Model APIs for commercial, open-source, and custom models, and Genie for natural-language queries over governed data; the intended result is faster experimentation and deployment with visibility into access, spending, lineage, and auditability.
Autonomous AI agents depend on the accuracy of the underlying data layer because retrieved data populates the LLM context window for multi-step reasoning. In distributed architectures, standard asynchronous replication lag causes agents to ingest stale data, leading to flawed execution and self-reinforcing hallucination debt when errors are saved back to memory. To resolve this, architectures must align replication strategies with specific task truth requirements across three primary patterns. High-stakes workloads utilize Amazon Aurora with global write forwarding or native Aurora DSQL for multi-region strong consistency, while high-scale scenarios leverage DynamoDB Global Tables with conditional writes to prevent lost updates. Finally, high-velocity ingestion pipelines enforce LOCAL_QUORUM reads and writes in Amazon Keyspaces to capture critical data spikes reliably.
The post presents a constraint-aware GPU allocator as a response to utilization losses caused by fixed real-time reservations and FIFO ordering under contention. It compares the allocator with FIFO on identical hardware and workloads across seven benchmark scenarios, modeling batch-like jobs as contiguous GPU blocks and real-time inference as a timestep-by-timestep demand curve with bounded GPU swaps. Across five contention scenarios, utilization improved from 52–85% to 72–88%, while priority-weighted value increased 24.6–105.1%; the strongest case gained 33 percentage points of utilization and 105% value, and the scale test gained 15.9% value at identical utilization and throughput. The scheduler optimizes a 24-hour horizon but commits only the current timestep, re-running every 30–60 minutes so fresh observations absorb forecast error while preserving running work.
Adobe Firefly powers generative creative features using a large-scale GPU training infrastructure hosted on Amazon Elastic Kubernetes Service. As workloads scaled to thousands of nodes and GPUs, Adobe's self-managed Prometheus setup struggled with query latency across high-cardinality telemetry data. To resolve these performance bottlenecks, the team moved two million critical time series metrics to Amazon Managed Service for Prometheus while using managed scrapers alongside their existing infrastructure. This setup accelerated query response times by more than 28x over 24-hour windows, reducing query execution times to approximately ten seconds. Additionally, the migration expanded practical monitoring windows from six hours to 24 hours, giving teams reliable visibility across long-running distributed training jobs.
Databricks’ serverless platform launches tens of millions of VMs daily, and each VM needs network configuration before serving customer workloads. The original design synchronously aggregated data from multiple upstream services on the cluster-creation critical path, producing 5,000-ms p99 RPC latency and availability bottlenecks as request volume reached billions per day. The replacement uses change events through a message queue, asynchronous per-workspace recomputation, partition-local snapshots, and periodic reconciliation to recover from missed events. The serving path now performs a single snapshot-store read without upstream calls, while events carry workspace and resource identifiers rather than customer data. After rollout, p99 RPC latency fell to 125 ms, server success rate rose to 99.99%, upstream call volume dropped 86%, and the legacy synchronous framework was deprecated.
Alibaba’s Qwen3.8-2.4T-A95B (Qwen3.8-Max) is an open-weight 2.4T-parameter model with 95B activated per token, designed for reasoning and agentic workloads. Its fine-grained mixture-of-experts architecture combines full and linear attention, supporting context of up to one million tokens and output lengths up to 128K while using bounded recurrent state in linear-attention layers. NVIDIA describes multinode serving on the 72-GPU GB300 NVL72, where a 130 TB/s NVLink domain supports expert communication; in FP8, the model reportedly exceeds 4K tokens per second per GPU and 350 tokens per second per user without additional tuning. Developers can control reasoning depth with low, high, or xhigh modes, choose SGLang, vLLM, NVIDIA Dynamo, or NIM for serving, and use NeMo AutoModel for full SFT or LoRA post-training.
Traditional multi-agent architectures rely on a central supervisor agent to decompose tasks and aggregate results, creating single points of failure and context-window throughput bottlenecks. To address these scaling limits on decomposable workloads, kiro-flock implements self-organizing clusters of headless Kiro CLI agents on Amazon EC2 coordinated entirely through shared Amazon S3 state. Agents operate in bounded peer topologies such as logical rings, reading neighbor append-only logs and generating artifacts without direct inter-agent messaging or central brokers. Each agent starts fresh per iteration without conversational memory, reading current direction and peer logs to prevent drift and preserve steering control. This shared-state pattern enables resilient parallelism for tasks requiring diverse independent contributions, such as code reviews or codebase migrations.
Salesforce faced scaling bottlenecks when individual engineering teams maintained fragmented, custom telemetry pipelines requiring manual data transformations and dashboard creation every release. To resolve these silos, the engineering organization established the Product Data Platform (PDP) using a standardized custom schema built atop Monitoring Cloud infrastructure. This framework mandates core telemetry fields for consistent metric generation while offering governed optional attributes for product-specific flexibility. An AI-driven Model Context Protocol tool assists developers by recommending compliant instrumentation patterns directly within their workflows. Today, the unified platform processes 45 billion rows of data daily across 19,000 distinct events and 2,000 product features, reducing dashboard refresh latency by 97 percent from one month to daily updates.
Platforms using blended payment rates absorb interchange variance, which can reduce margins as card mix shifts toward premium rewards, corporate, and other high-interchange cards. Network cost passthrough, also called IC++, lets eligible Connect platforms pass interchange, scheme fees, and applicable adjustments to connected accounts at cost while keeping Stripe fees and platform markup separate. Eligibility requires IC+ pricing, platform-controlled pricing, enabled transfers, an appropriate Merchant of Record fund flow, and eligible geographies; passthrough can then be enabled in the Dashboard or through the API. The rollout also requires lower application fees, new network-cost reporting, and a pilot with clear communication before broader deployment.
Musinsa’s Audience Engine powers a customer data platform that targets audiences across 16.4 million members. It stores each user-to-audience relationship as a separate row, producing roughly 1.1 billion cohort-to-user mappings across 2,500 audiences and potentially 16 million new rows when one audience is added. Its self-hosted ClickHouse deployment coupled storage and compute, forced workloads to share a cluster, and created operational burdens as data and query demands grew. Musinsa migrated to ClickHouse Cloud on AWS, using separated S3-backed storage, workload-specific compute, automatic idling, vertical autoscaling, and ClickPipes for direct real-time ingestion. The company reports an 86.5% reduction in storage costs and a total cost of ownership reduction of up to 71.4%, while shifting engineering effort toward CDP business logic.
Netflix designed a query execution serving layer for its Real-Time Distributed Graph to power sub-100ms responses across diverse graph traversal workloads spanning billions of nodes and edges. To prevent compounding network delays in multi-hop queries, the engine uses a breadth-first traversal model that batches entity lookups across entire frontier levels rather than tracing paths depth-first. The entire serving architecture relies on asynchronous composition across small thread pools of 16 to 24 threads, ensuring no thread blocks while waiting on remote storage or enrichment I/O. Selective caching via EVCache targets stable properties with volatility-matched TTLs, producing 70 to 80 percent cache hit rates and reducing backend storage calls by three to four times. A layered filtering hierarchy pushes depth and edge limits directly to the storage tier, eliminating bespoke code changes while isolating system resources against excessive fan-out.
Salesforce redesigned its internal observability platform, Argus, to eliminate single-region failure risks while ingesting approximately 4 billion metrics per minute. Operating out of a single AWS region previously created a global blast radius and generated high cross-region data transfer costs. To mitigate these risks without the prohibitive expense of full multi-region replication, the engineering team implemented a geo-local architecture that processes and stores telemetry closer to its origin across production geographies. A newly introduced federation query layer relies on Elasticsearch mappings to route queries selectively to relevant regional clusters, avoiding broad fan-out. The platform also adopted metadata caching for wildcard queries and implemented HTTP 206 partial-response handling to maintain visibility when individual regions experience downtime.
ClickHouse Cloud’s vertical autoscaling recommender originally ran on a fixed timer, creating latency when workloads suddenly increased and services needed resources before the next scheduled pass. The redesign keeps the periodic sweep as a safety net while adding a reactive fast path, using controller-runtime as an event-processing layer with sources feeding a shared rate-limited work queue and reconciler. The queue provides deduplication, exponential backoff, bounded concurrency, leader election, and metrics, while ClickHouse stores near-real-time signals that a polling source reads; the excerpt notes polling sets a floor on reaction speed. The reactive path is rolling out cautiously in observe-only mode, with periodic reconciliation remaining the main driver, and the stated goal is to trigger scale-ups for critical events such as out-of-memory conditions in seconds rather than waiting for the next scheduled run.
Enterprise AI adoption across regulated industries stalls when internal governance mandates private, dedicated network paths for sensitive data. To eliminate the need for manual VPN tunnels and complex proxy layers, Salesforce re-engineered Private Connect into a multi-cloud networking platform connecting customer environments to Hyperforce services. The Private Connect v2.0 architecture utilizes private links, transit gateways, and direct endpoint routing alongside an automated control plane for rapid Hyperforce regional rollouts. In addition to supporting cross-substrate interconnects for Azure and decoupling endpoint management, the platform incorporates a generic connector framework supporting tools like Snowflake, Databricks, and Kafka. Today, the system provisions secure connectivity in under 30 minutes while processing roughly 120 TB of data and 683 million requests monthly across 15 AWS regions.
Agentic AI architectures require storage systems to secure, compress, and validate massive streams of persistent data and key-value cache entries across concurrent reasoning loops. These tasks run primarily on CPUs and frequently create processing bottlenecks that prevent modern networks and flash storage from operating at full capacity. NVIDIA evaluated the Vera CPU within the BlueField-4 STX Storage Processor using a purpose-built in-memory microbenchmark suite running standard libraries like OpenSSL, Zstandard, and LZ4. The results demonstrated that Vera outperformed x86 processors across core storage primitives including encryption, decompression, Reed-Solomon recovery, and CRC32C integrity checking. For a sequential multi-stage pipeline executing compression followed by encryption, Vera achieved up to 3.21x higher throughput than the comparison x86 CPU.
As context lengths increase in long-context and agentic workloads, attention increasingly dominates GPU inference runtime. Dense attention performance depends on architectural parameters like group size, head dimension, and sequence length, alongside GPU parallelization strategies. Prefill remains compute-bound and scales quadratically with input sequence length, showing minimal runtime sensitivity to group size variations. Conversely, decode generation is memory-bound and scales linearly with KV cache length, where doubling group size roughly halves decode execution time. To maximize throughput and interactivity on NVIDIA hardware, model architects should push group sizes higher, select head dimensions of 128 or 256, reduce effective KV state, and keep tensor parallelism less than or equal to the number of KV heads.
Data organizations face severe operational risks from inconsistent metrics, sprawling access controls, and incomplete change management across disparate BI tools and warehouses. Traditional remediation relies on centralized BI gatekeepers and fragmented governance frameworks, which introduces costly bottlenecks and fails to scale. Implementing a semantic layer mitigates these risks by consolidating metric definitions, business logic, and calculations into a single governed location. Centralizing access shrinks the governance surface area and allows metric updates to propagate automatically to downstream consumers, including Tableau, Power BI, Excel, and AI tools. Structured metadata captures context directly alongside models, enabling reliable self-service and federated hub-and-spoke delivery while keeping definitions auditable under version control.
The best engineering reads, once a week.
