[{"data":1,"prerenderedAt":1664},["ShallowReactive",2],{"active-banner":3,"navbar-featured-partner-blog":23,"navbar-pricing-featured":304,"blog-\u002Fblog\u002Fhow-pulsars-architecture-delivers-better-performance-than-kafka":1084,"blog-authors-\u002Fblog\u002Fhow-pulsars-architecture-delivers-better-performance-than-kafka":1616,"related-\u002Fblog\u002Fhow-pulsars-architecture-delivers-better-performance-than-kafka":1642},{"id":4,"title":5,"date":6,"dismissible":7,"extension":8,"link":9,"link2":10,"linkText":11,"linkText2":10,"meta":12,"stem":20,"variant":21,"__hash__":22},"banners\u002Fbanners\u002Fkafka-company-2025.md","Native Apache Kafka Service Is Coming Soon to StreamNative Cloud. Join the waitlist and get $1,000 in credits.","2026-04-01",true,"md","\u002Fnative-kafka-service-waitlist",null,"Join Waitlist",{"body":13},{"type":14,"value":15,"toc":16},"minimark",[],{"title":17,"searchDepth":18,"depth":18,"links":19},"",2,[],"banners\u002Fkafka-company-2025","default","IMIJszQOOWTfA_DV33eYUA5jqV7DrX1FWbBTBZfNvWc",{"id":24,"title":25,"authors":26,"body":28,"category":288,"createdAt":10,"date":289,"description":290,"extension":8,"featured":7,"image":291,"isDraft":292,"link":10,"meta":293,"navigation":7,"order":294,"path":295,"readingTime":296,"relatedResources":10,"seo":297,"stem":298,"tags":299,"__hash__":303},"blogs\u002Fblog\u002Fstreamnative-recognized-in-the-forrester-wave-streaming-data-platforms-2025.md","StreamNative Recognized as a Contender in The Forrester Wave™: Streaming Data Platforms, Q4 2025",[27],"David Kjerrumgaard",{"type":14,"value":29,"toc":275},[30,38,46,50,66,72,77,80,86,101,108,114,117,123,126,133,139,142,145,156,162,168,171,174,177,183,190,193,196,203,206,209,223,228,232,236,240,244,248,250,267,269],[31,32,34],"h3",{"id":33},"receives-highest-possible-scores-in-both-the-messaging-and-resource-optimization-criteria",[35,36,37],"em",{},"Receives Highest Possible Scores in BOTH the Messaging and Resource Optimization Criteria",[39,40,42],"h2",{"id":41},"introduction",[43,44,45],"strong",{},"Introduction",[47,48,49],"p",{},"Real-time data has become the backbone of modern innovation. As artificial intelligence (AI) and digital services demand instantaneous insights, organizations are realizing that streaming data is no longer optional – it's essential for delivering timely, context-rich experiences. StreamNative's data streaming platform is built precisely for this reality, ensuring data is immediate, reliable, and ready to power critical applications.",[47,51,52,53,62,63],{},"Today, we're excited to announce that Forrester Research has named StreamNative as a Contender in its evaluation, ",[54,55,57],"a",{"href":56},"\u002Freports\u002Frecognized-in-the-forrester-wave-tm-streaming-data-platforms-q4-2025",[35,58,59],{},[43,60,61],{},"The Forrester Wave™: Streaming Data Platforms, Q4 2025",". This report evaluated 15 top streaming data platform providers, and we're proud to share that ",[43,64,65],{},"StreamNative received the highest scores possible—5 out of 5—in both the Messaging and Resource Optimization criteria.",[47,67,68,69],{},"***Forrester's Take: ***",[35,70,71],{},"\"StreamNative is a good fit for enterprises that want an Apache Pulsar implementation that is also compatible with Kafka APIs.\"",[47,73,74],{},[35,75,76],{},"— The Forrester Wave™: Streaming Data Platforms, Q4 2025",[47,78,79],{},"Being recognized in the Forrester Wave is a proud milestone, and for us, it highlights how far StreamNative has come in enabling enterprises to unlock the power of real-time data. In the sections below, we'll dive into what we believe sets StreamNative apart—from our modern architecture and cloud-native design to our open-source foundation and real-time use cases—and how we see these strengths aligning with Forrester's findings.",[39,81,83],{"id":82},"trusted-by-industry-leaders",[43,84,85],{},"Trusted by Industry Leaders",[47,87,88,89,92,93,96,97,100],{},"Companies across industries are already leveraging StreamNative to drive real-time outcomes. Global enterprises like ",[43,90,91],{},"Cisco"," rely on StreamNative to handle massive IoT telemetry, supporting 245 million+ connected devices. Martech leaders such as ",[43,94,95],{},"Iterable"," process billions of events per day with StreamNative for hyper-personalized customer engagement. And in financial services, ",[43,98,99],{},"FICO"," trusts StreamNative to power its real-time fraud detection and analytics pipelines with a secure, scalable streaming backbone.",[47,102,103,104,107],{},"The Forrester report notes that, “",[35,105,106],{},"Customers appreciate the lower infrastructure costs that result from StreamNative’s cost-efficient, Kafka-compatible architecture. Customers note excellent support responsiveness…","”",[39,109,111],{"id":110},"modern-cloud-native-architecture-built-for-scale",[43,112,113],{},"Modern, Cloud-Native Architecture Built for Scale",[47,115,116],{},"From day one, StreamNative was designed with a modern architecture to meet the demanding scale and flexibility requirements of real-time data. Unlike legacy streaming systems that often rely on tightly coupled storage and compute, StreamNative's platform takes a cloud-native approach: it decouples these layers to enable elastic scalability and efficient resource utilization across any environment. The core is powered by Apache Pulsar—a distributed messaging and streaming engine—enhanced with multi-protocol support (including native Apache Kafka API compatibility) to unify diverse data streams under one roof. This means organizations can consolidate siloed messaging systems and handle both high-volume event streams and traditional message queues on a single platform, without sacrificing performance or reliability.",[47,118,119,120,107],{},"Forrester's evaluation described that “",[35,121,122],{},"StreamNative aims to provide a high-performance, multi-protocol streaming data platform: It uses Apache Pulsar with Kafka API compatibility to deliver cost-efficient, real-time applications for enterprises. It appeals to organizations that want a flexible, low-cost streaming solution, due to its focus on scalability and resource optimization, while its investments in Pulsar’s open-source ecosystem and performance optimization make it the primary platform for enterprises wishing to implement Pulsar.",[47,124,125],{},"Our cloud-first, leaderless architecture (with no single broker bottlenecks) and tiered storage model were built to maximize throughput and cost-efficiency for real-time workloads. By separating compute from storage and leveraging distributed object storage, StreamNative can retain huge volumes of event data indefinitely while keeping compute costs in check—effectively providing a flexible, low-cost streaming solution.",[47,127,128,129,132],{},"This modern design not only delivers high performance, but also ensures fault tolerance and geo-distribution out of the box, so enterprises can trust their streaming data is always available and durable. As Forrester’s evaluation noted, StreamNative ",[35,130,131],{},"\"excels at messaging and resource optimization\" and “Its platform supports use cases like real-time analytics and event-driven architectures with robust scalability.","” Our architecture provides the strong foundation that today's real-time applications demand, from ultra-fast data ingestion to seamless scale-out across hybrid and multi-cloud environments.",[39,134,136],{"id":135},"open-source-foundation-and-pulsar-expertise",[43,137,138],{},"Open Source Foundation and Pulsar Expertise",[47,140,141],{},"StreamNative's DNA is rooted in open source innovation. Our founders are the original creators of Apache Pulsar, and we've built our platform with the same open principles: freedom, flexibility, and community-driven innovation. For developers and data teams, this means adopting StreamNative comes with no proprietary lock-in—instead, you get a platform built on open standards and a thriving ecosystem. We offer broad API compatibility (Pulsar, Kafka, JMS, MQTT, and more) so that teams can work with familiar interfaces and integrate StreamNative into existing systems with ease.",[47,143,144],{},"StreamNative is the primary commercial contributor to the Apache Pulsar project and its surrounding ecosystem. We invest heavily in Pulsar's ongoing improvements our investments in Pulsar's open-source ecosystem and performance optimization bolster StreamNative's value. We also foster a vibrant community through initiatives like the Data Streaming Summit and free training resources.",[47,146,147,148,151,152,155],{},"Forrester's assessment noted that StreamNative’s “",[35,149,150],{},"events-driven agents, extensibility, and performance architecture are solid,","” and we're continuing to build on that foundation. ",[43,153,154],{},"We're actively investing in expanding our tooling for observability, governance, schema management, and developer productivity","—areas we recognize as critical for enterprise adoption and where we're committed to accelerating our roadmap.",[47,157,158,159],{},"Being open also means embracing an open ecosystem of technologies. StreamNative actively integrates with the tools and platforms that matter most to our users. We partner with industry leaders like Snowflake, Databricks, Google, and Ververica to ensure our streaming platform works seamlessly with data warehouses, lakehouse storage, and stream processing frameworks. Forrester’s evaluation observed that StreamNative’s ",[35,160,161],{},"\"investments in Pulsar’s open-source ecosystem and performance optimization make it the primary platform for enterprises wishing to implement Pulsar.\"",[39,163,165],{"id":164},"powering-real-time-use-cases-across-industries",[43,166,167],{},"Powering Real-Time Use Cases Across Industries",[47,169,170],{},"One of the greatest validations of StreamNative's approach is the success our customers are achieving with real-time data. StreamNative's platform is versatile and use-case agnostic—if an application demands high-volume, low-latency data movement, we can power it. This flexibility is why our customer base spans industries from finance and IoT to major automobile manufacturers and online gaming. The common thread is that these organizations need to process and react to data in milliseconds, and StreamNative is delivering the capabilities to make that possible.",[47,172,173],{},"Cisco uses StreamNative to underpin an IoT telemetry system of colossal scale, connecting hundreds of millions of devices and thousands of enterprise clients with real-time data streams. The platform's multi-tenant design and proven reliability allow Cisco to offer its customers a live feed of device data with unwavering confidence. In the financial sector, FICO has built streaming pipelines on StreamNative to detect fraud as transactions happen and to monitor systems in real time. With StreamNative's strong guarantees around message durability and ordering, FICO can catch anomalies or suspicious patterns within seconds. And in digital customer engagement, Iterable relies on StreamNative to process billions of events every day—clicks, views, purchases—so that marketers can trigger personalized campaigns instantly based on user behavior.",[47,175,176],{},"Our customers uniformly deal with mission-critical data streams, where downtime or delays are unacceptable. StreamNative's fault-tolerant, scalable infrastructure has proven equal to the task, handling scenarios like bursting to millions of events per second or seamlessly spanning multiple cloud regions. Forrester's report recognized StreamNative for supporting event-driven architectures with robust scalability—which for us is a reflection of our platform's ability to meet the most demanding enterprise requirements.",[39,178,180],{"id":179},"continuing-to-innovate-ursa-orca-and-the-road-ahead",[43,181,182],{},"Continuing to Innovate: Ursa, Orca, and the Road Ahead",[47,184,185,186,189],{},"While we are thrilled to be recognized in Forrester's Streaming Data Platforms Wave, we view this as just the beginning. StreamNative's vision has always been bold: to ",[43,187,188],{},"provide a unified platform that not only handles today's streaming needs but also anticipates the emerging requirements of tomorrow",".",[47,191,192],{},"One key area of focus is the convergence of streaming data with advanced analytics and AI. As Forrester points out in the report, technology leaders should look for platforms that natively integrate messaging, stream processing, and analytics to provide AI agents with real-time, contextualized information. We couldn't agree more. Our award-winning Ursa Engine and Orca Agent Engine are aimed at extending our platform up the stack—bridging the gap between data streams and data lakes, and between event streams and intelligent processing.",[47,194,195],{},"Our new Ursa Engine introduces a lakehouse-native approach to streaming: it can write events directly to table formats like Iceberg on cloud storage, eliminating entire classes of ETL jobs and making fresh data instantly available for analytics queries. By integrating streaming and lakehouse technologies, we help customers collapse data silos and accelerate their AI\u002FML pipelines.",[47,197,198,199,202],{},"Beyond analytics integration, we are also enhancing StreamNative with more out-of-the-box processing and governance capabilities. In the coming months, we plan to introduce new features for lightweight stream processing and transformation, making it easier to build reactive applications directly on the platform. We're also expanding our ecosystem of connectors and integrations, so that whether your data lands in Snowflake, Databricks, or an AI model, StreamNative will seamlessly feed it. ",[43,200,201],{},"We're investing significantly in enterprise features including security, schema registry, governance, and monitoring tooling","—capabilities that are essential for mission-critical deployments and where we're committed to continued improvement.",[47,204,205],{},"This recognition from Forrester energizes us to keep innovating at full speed. We're sharing this honor with our amazing customers, community, and partners who drive us forward every day. Your feedback and real-world challenges have helped shape StreamNative into what it is today, and together, we will shape the future of streaming data. Thank you for joining us on this journey—we're just getting started, and we can't wait to deliver even more value as we continue to evolve our platform. Onward to real-time everything!",[207,208],"hr",{},[31,210,212],{"id":211},"streamnative-in-the-forrester-wave-evaluation-findings",[43,213,214,215,222],{},"StreamNative in ",[43,216,217],{},[54,218,219],{"href":56},[43,220,221],{},"The Forrester Wave™",": Evaluation Findings",[224,225,227],"h5",{"id":226},"recognized-as-a-contender-among-15-streaming-data-platform-providers","• Recognized as a Contender among 15 streaming data platform providers",[224,229,231],{"id":230},"received-the-highest-scores-possible-50-in-both-the-messaging-and-resource-optimization-criteria","* Received the highest scores possible (5.0) in both the Messaging and Resource Optimization criteria",[224,233,235],{"id":234},"cited-as-the-primary-platform-for-enterprises-wishing-to-implement-pulsar","• Cited as the primary platform for enterprises wishing to implement Pulsar",[224,237,239],{"id":238},"noted-for-excelling-at-messaging-and-resource-optimization","• Noted for excelling at messaging and resource optimization",[224,241,243],{"id":242},"customers-cited-lower-infrastructure-costs-and-excellent-support-responsiveness","• Customers cited lower infrastructure costs and excellent support responsiveness",[224,245,247],{"id":246},"recognized-for-supporting-event-driven-architectures-with-robust-scalability","• Recognized for supporting event-driven architectures with robust scalability",[207,249],{},[251,252,254,255,258,259,189],"h6",{"id":253},"forrester-disclaimer-forrester-does-not-endorse-any-company-product-brand-or-service-included-in-its-research-publications-and-does-not-advise-any-person-to-select-the-products-or-services-of-any-company-or-brand-based-on-the-ratings-included-in-such-publications-information-is-based-on-the-best-available-resources-opinions-reflect-judgment-at-the-time-and-are-subject-to-change-for-more-information-read-about-forresters-objectivity-here","**Forrester Disclaimer: **",[35,256,257],{},"Forrester does not endorse any company, product, brand, or service included in its research publications and does not advise any person to select the products or services of any company or brand based on the ratings included in such publications. Information is based on the best available resources. Opinions reflect judgment at the time and are subject to change",". *For more information, read about Forrester’s objectivity *",[54,260,264],{"href":261,"rel":262},"https:\u002F\u002Fwww.forrester.com\u002Fabout-us\u002Fobjectivity\u002F",[263],"nofollow",[35,265,266],{},"here",[207,268],{},[251,270,272],{"id":271},"apache-apache-pulsar-apache-kafka-apache-flink-and-other-names-are-trademarks-of-the-apache-software-foundation-no-endorsement-by-apache-or-other-third-parties-is-implied",[35,273,274],{},"Apache®, Apache Pulsar®, Apache Kafka®, Apache Flink® and other names are trademarks of The Apache Software Foundation. No endorsement by Apache or other third parties is implied.",{"title":17,"searchDepth":18,"depth":18,"links":276},[277,279,280,281,282,283,284],{"id":33,"depth":278,"text":37},3,{"id":41,"depth":18,"text":45},{"id":82,"depth":18,"text":85},{"id":110,"depth":18,"text":113},{"id":135,"depth":18,"text":138},{"id":164,"depth":18,"text":167},{"id":179,"depth":18,"text":182,"children":285},[286],{"id":211,"depth":278,"text":287},"StreamNative in The Forrester Wave™: Evaluation Findings","Company","2025-12-16","StreamNative is recognized in The Forrester Wave™: Streaming Data Platforms, Q4 2025. Discover why Forrester highlights StreamNative's high-performance messaging, efficient resource use, and cost-effective Kafka API compatibility for real-time innovation.","\u002Fimgs\u002Fblogs\u002F693bd36cf01b217dcb67278f_Streamnative_blog_thumbnail.png",false,{},0,"\u002Fblog\u002Fstreamnative-recognized-in-the-forrester-wave-streaming-data-platforms-2025","10 mins read",{"title":25,"description":290},"blog\u002Fstreamnative-recognized-in-the-forrester-wave-streaming-data-platforms-2025",[300,301,302],"Announcements","Real-Time","Forrester","sOeeJtEO3O-IIfTPJjY1AFOMawZ_rf8FOH8A98NEKgU",{"id":305,"title":306,"authors":307,"body":312,"category":1071,"createdAt":10,"date":1072,"description":1073,"extension":8,"featured":7,"image":1074,"isDraft":292,"link":10,"meta":1075,"navigation":7,"order":294,"path":1076,"readingTime":1077,"relatedResources":10,"seo":1078,"stem":1079,"tags":1080,"__hash__":1083},"blogs\u002Fblog\u002Fhow-we-run-a-5-gb-s-kafka-workload-for-just-50-per-hour.md","How We Run a 5 GB\u002Fs Kafka Workload for Just $50 per Hour",[308,309,310,311],"Matteo Meril","Neng Lu","Hang Chen","Penghui Li",{"type":14,"value":313,"toc":1041},[314,317,320,323,326,329,333,336,346,352,355,363,368,372,379,382,385,393,397,400,405,409,412,415,418,421,430,434,437,448,451,455,458,461,472,475,479,483,491,494,498,506,535,539,542,547,551,554,558,561,564,569,578,583,586,589,600,604,607,618,622,625,628,633,636,665,669,671,677,680,685,690,693,697,711,715,726,730,745,754,765,768,771,775,778,781,792,795,798,801,806,811,815,819,836,840,854,859,863,874,877,893,897,908,913,918,926,930,933,937,944,948,951,960,965,974,980,989,998,1007,1016,1025,1033],[47,315,316],{},"The rise of DeepSeek has shaken the AI infrastructure market, forcing companies to confront the escalating costs of training and deploying AI models. But the real pressure point isn’t just compute—it’s data acquisition and ingestion costs.",[47,318,319],{},"As businesses rethink their AI cost-containment strategies, real-time data streaming is emerging as a critical enabler. The growing adoption of Kafka as a standard protocol has expanded cost-efficient options, allowing companies to optimize streaming analytics while keeping expenses in check.",[47,321,322],{},"Ursa, the data streaming engine powering StreamNative’s managed Kafka service, is built for this new reality. With its leaderless architecture and native lakehouse storage integration, Ursa eliminates costly inter-zone network traffic for data replication and client-to-broker communication while ensuring high availability at minimal operational cost.",[47,324,325],{},"In this blog post, we benchmarked the infrastructure cost and total cost of ownership (TCO) for running a 5GB\u002Fs Kafka workload across different Kafka vendors, including Redpanda, Confluent WarpStream, and AWS MSK. Our benchmark results show that Ursa can sustain 5GB\u002Fs Kafka workloads at just 5% of the cost of traditional streaming engines like Redpanda—making it the ideal solution for high-performance, cost-efficient ingestion and data streaming for data lakehouses and AI workloads.",[47,327,328],{},"Note: We also evaluated vanilla Kafka in our benchmark; however, for simplicity, we have focused our cost comparison on vendor solutions rather than self-managed deployments. That said, it is important to highlight that both Redpanda and vanilla Kafka use a leader-based data replication approach. In a data-intensive, network-bound workload like 5GB\u002Fs streaming, with the same machine type and replication factor, Redpanda and vanilla Kafka produced nearly identical cost profiles.",[39,330,332],{"id":331},"key-benchmark-findings","Key Benchmark Findings",[47,334,335],{},"Ursa delivered 5 GB\u002Fs of sustained throughput at an infrastructure cost of just $54 per hour. For comparison:",[337,338,339,343],"ul",{},[340,341,342],"li",{},"MSK: $303 per hour → 5.6x more expensive compared to Ursa",[340,344,345],{},"Redpanda: $988 per hour → 18x more expensive compared to Ursa",[47,347,348],{},[349,350],"img",{"alt":17,"src":351},"\u002Fimgs\u002Fblogs\u002F679c71b67d9046f26edc7977_AD_4nXfvTqyBNUBu2lObdkKAx-5UNkpNP8UYULLZyOcixE6z99VMZUUEsUqWjzexI7vjyNGRNSAUoM9smYvdTP55ctAhIbrs5lmQgcSVMWdaoigbWouCl95DVSQsxooY-qqfGcYqS4g4zA.png",[47,353,354],{},"Beyond infrastructure costs, when factoring in both storage pricing, vendor pricing and operational expenses, Ursa’s total cost of ownership (TCO) for a 5GB\u002Fs workload with a 7-day retention period is:",[337,356,357,360],{},[340,358,359],{},"50% cheaper than Confluent WarpStream",[340,361,362],{},"85% cheaper than MSK and Redpanda",[47,364,365],{},[349,366],{"alt":17,"src":367},"\u002Fimgs\u002Fblogs\u002F679c602d77e9c706de5343b8_AD_4nXeDv8rrv_C1CTCCiqYo1zpvlGYbdBk1r0VEqovAPu22iFMQZgh54Hfw9PBMLzM7jDFxKwAFDxbdG0np4XVk_tGsWhEKMloLRcmmea7lvueCx-0cFsyaE3Mya4Mxc1Dox95A6JEc.png",[39,369,371],{"id":370},"ursa-highly-cost-efficient-data-streaming-at-scale","Ursa: Highly Cost-Efficient Data Streaming at Scale",[47,373,374,378],{},[54,375,377],{"href":376},"\u002Fblog\u002Fursa-reimagine-apache-kafka-for-the-cost-conscious-data-streaming","Ursa"," is a next-generation data streaming engine designed to deliver high performance at a fraction of the cost of traditional disk-based solutions. It is fully compatible with Apache Kafka and Apache Pulsar APIs, while leveraging a leaderless, lakehouse-native architecture to maximize scalability, efficiency, and cost savings.",[47,380,381],{},"Ursa’s key innovation is separating storage from compute and decoupling metadata\u002Findex operations from data operations by utilizing cloud object storage (e.g., AWS S3) instead of costly inter-zone disk-based replication. It also employs open lakehouse formats (Iceberg and Delta Lake), enabling columnar compression to significantly reduce storage costs while maintaining durability and availability.",[47,383,384],{},"In contrast, traditional streaming systems—like Kafka and Redpanda—depend on leader-based architectures, which drive up inter-zone traffic costs due to replication and client communication. Ursa mitigates these costs by:",[337,386,387,390],{},[340,388,389],{},"Eliminating inter-zone traffic costs via a leaderless architecture.",[340,391,392],{},"Replacing costly inter-zone replication with direct writes to cloud storage using open lakehouse formats.",[39,394,396],{"id":395},"how-ursa-eliminates-inter-zone-traffic","How Ursa Eliminates Inter-Zone Traffic",[47,398,399],{},"Ursa minimizes inter-zone traffic by leveraging a leaderless architecture, which eliminates inter-zone communication between clients and brokers, and lakehouse-native storage, which removes the need for inter-zone data replication. This approach ensures high availability and scalability while avoiding unnecessary cross-zone data movement.",[47,401,402],{},[349,403],{"alt":17,"src":404},"\u002Fimgs\u002Fblogs\u002F679c602e21b3571bb7117dca_AD_4nXd7Oahc77NjRLNvA9clLt0tsyU6MrIqVibFYv5pW5giTIcCHPr3EA_yTGzfVEUIVO3VXK56qWK8zmBCp5lY0E_4nmlWIPFrHjtHylA5NhwELjn-UB0fLG2h_kbrxrc7Cs_edvveNA.png",[31,406,408],{"id":407},"leaderless-architecture","Leaderless architecture",[47,410,411],{},"Traditional streaming engines such as Kafka, Pulsar, or RedPanda rely on a leader-based model, where each partition is assigned to a single leader broker that handles all writes and reads.",[47,413,414],{},"Pros of Leader-Based Architectures:\n✔ Maintains message ordering via local sequence IDs\n✔ Delivers low latency and high performance through message caching",[47,416,417],{},"Cons of Leader-Based Architectures:\n✖ Throughput bottlenecked by a single broker per partition\n✖ Inter-zone traffic required for high availability in multi-AZ deployments",[47,419,420],{},"While Kafka and Pulsar offer partial solutions (e.g., reading from followers, shadow topics) to reduce read-related inter-zone traffic, producers still send data to a single leader.",[47,422,423,424,429],{},"Ursa removes the concept of topic ownership, allowing any broker in the cluster to handle reads or writes for any partition. The primary challenge—ensuring message ordering—is solved with ",[54,425,428],{"href":426,"rel":427},"https:\u002F\u002Fgithub.com\u002Fstreamnative\u002Foxia",[263],"Oxia",", a scalable metadata and index service created by StreamNative in 2022.",[31,431,433],{"id":432},"oxia-the-metadata-layer-enabling-leaderless-architecture","Oxia: The Metadata Layer Enabling Leaderless Architecture",[47,435,436],{},"Ensuring message ordering in a leaderless architecture is complex, but Ursa solves this with Oxia:",[337,438,439,442,445],{},[340,440,441],{},"Handles millions of metadata\u002Findex operations per second",[340,443,444],{},"Generates sequential IDs to maintain strict message ordering",[340,446,447],{},"Optimized for Kubernetes with horizontal scalability",[47,449,450],{},"Producers and consumers can connect to any broker within their local AZ, eliminating inter-zone traffic costs while maintaining performance through localized caching.",[31,452,454],{"id":453},"zero-interzone-data-replication","Zero interzone data replication",[47,456,457],{},"In most distributed systems, data replication from a leader (primary) to followers (replicas) is crucial for fault tolerance and availability. However, replication across zones can inflate infrastructure expenses substantially.",[47,459,460],{},"Ursa avoids these costs by writing data directly to cloud storage (e.g., AWS S3, Google GCS):",[337,462,463,466,469],{},[340,464,465],{},"Built-In Resilience: Cloud storage inherently offers high availability and fault tolerance without inter-zone traffic fees.",[340,467,468],{},"Tradeoff: Slightly higher latency (sub-second, with p99 at 500 milliseconds) compared to local disk\u002FEBS (single-digit to sub-100 milliseconds), in exchange for significantly lower costs (up to 10x lower).",[340,470,471],{},"Flexible Modes: Ursa is an addition to the classic BookKeeper-based engine, providing users with the flexibility to optimize for either cost or low latency based on their workload requirements.",[47,473,474],{},"By foregoing conventional replication, Ursa slashes inter-zone traffic costs and associated complexities—making it a compelling option for organizations seeking to balance high-performance data streaming with strict budget constraints.",[39,476,478],{"id":477},"how-we-ran-a-5-gbs-test-with-ursa","How We Ran a 5 GB\u002Fs Test with Ursa",[31,480,482],{"id":481},"ursa-cluster-deployment","Ursa Cluster Deployment",[337,484,485,488],{},[340,486,487],{},"9 brokers across 3 availability zones, each on m6i.8xlarge (Fixed 12.5 Gbps bandwidth, 32 vCPU cores, 128 GB memory).",[340,489,490],{},"Oxia cluster (metadata store) with 3 nodes of m6i.8xlarge, distributed across three availability zones (AZs).",[47,492,493],{},"During peak throughput (5 GB\u002Fs), each broker’s network usage was about 10 Gbps.",[31,495,497],{"id":496},"openmessaging-benchmark-workers-configuration","OpenMessaging Benchmark Workers & Configuration",[47,499,500,501,505],{},"The OpenMessaging Benchmark(OMB) Framework is a suite of tools that make it easy to benchmark distributed messaging systems in the cloud. Please check ",[54,502,503],{"href":503,"rel":504},"https:\u002F\u002Fopenmessaging.cloud\u002Fdocs\u002Fbenchmarks\u002F",[263]," for details.",[337,507,508,523,532],{},[340,509,510,511,516,517,522],{},"12 OMB workers: 6 for ",[54,512,515],{"href":513,"rel":514},"https:\u002F\u002Fgist.github.com\u002Fcodelipenghui\u002Fd1094122270775e4f1580947f80c5055",[263],"producers",", 6 for ",[54,518,521],{"href":519,"rel":520},"https:\u002F\u002Fgist.github.com\u002Fcodelipenghui\u002F06bada89381fb77a7862e1b4c1d8963d",[263],"consumers"," across 3 availability zones, on m6i.8xlarge instances. Each worker is configured with 12 CPU cores and 48 GB memory.",[340,524,525,526,531],{},"Sample YAML ",[54,527,530],{"href":528,"rel":529},"https:\u002F\u002Fgist.github.com\u002Fcodelipenghui\u002F204c1f26c4d44a218ae235bf2de99904",[263],"scripts"," provided for Kafka-compatible configuration and rate limits.",[340,533,534],{},"Achieved consistent 5 GB\u002Fs publish\u002Fsubscribe throughput.",[39,536,538],{"id":537},"ursa-benchmark-tests-results","Ursa Benchmark Tests & Results",[47,540,541],{},"The following diagram demonstrates that Ursa can consistently handle 5 GB\u002Fs of traffic, fully saturating the network across all broker nodes.",[47,543,544],{},[349,545],{"alt":17,"src":546},"\u002Fimgs\u002Fblogs\u002F679c602d7b261bac1113f7d6_AD_4nXdDPsRc3koXICiFF0bqSmGWbJt_RlUy4FE3ruuWOfbCfpcqZ1dejjqGbkaCJv2hQFL1nirRouBVRW2l5uMWBvY9naMqGB_wHcLI14dBM0f85TXhmdm3UxEv1yGX9Y4hf5FttSkZew.png",[39,548,550],{"id":549},"comparing-infrastructure-cost","Comparing Infrastructure Cost",[47,552,553],{},"This benchmark first evaluates infrastructure costs of running a 5 GB\u002Fs streaming workload (1:1 producer-to-consumer ratio) across different data streaming engines, including Ursa, Redpanda, and AWS MSK, with a focus on multi-AZ deployments to ensure a fair comparison.",[31,555,557],{"id":556},"test-setup-key-assumptions","Test Setup & Key Assumptions",[47,559,560],{},"All tests use multi-AZ configurations, with clusters and clients distributed across three AWS availability zones (AZs). Cluster size scales proportionally to the number of AZs, and rack-awareness is enabled for all engines to evenly distribute topic partitions and leaders.",[47,562,563],{},"To ensure a fair comparison, we selected the same machine type capable of fully utilizing both network and storage bandwidth for Ursa and Redpanda in this 5GB\u002Fs test:",[337,565,566],{},[340,567,568],{},"9 × m6i.8xlarge instances",[47,570,571,572,577],{},"However, MSK's storage bandwidth limits vary depending on the selected instance type, with the highest allowed limit capped at 1000 MiB\u002Fs per broker, according to",[54,573,576],{"href":574,"rel":575},"https:\u002F\u002Fdocs.aws.amazon.com\u002Fmsk\u002Flatest\u002Fdeveloperguide\u002Fmsk-provision-throughput-management.html#throughput-bottlenecks",[263]," AWS documentation",". Given this constraint, achieving 5 GB\u002Fs throughput with a replication factor of 3 required the following setup:",[337,579,580],{},[340,581,582],{},"15 × kafka.m7g.8xlarge (32 vCPUs, 128 GB memory, 15 Gbps network, 4000 GiB EBS).",[47,584,585],{},"This configuration was necessary to work around MSK's storage bandwidth limitations, ensuring a comparable cost basis to other evaluated streaming engines.",[47,587,588],{},"Additional key assumptions include:",[337,590,591,594,597],{},[340,592,593],{},"Inter-AZ producer traffic: For leader-based engines, two-thirds of producer-to-broker traffic crosses AZs due to leader distribution.",[340,595,596],{},"Consumer optimizations: Follower fetch is enabled across all tests, eliminating inter-AZ consumer traffic.",[340,598,599],{},"Storage cost exclusions: This benchmark only evaluates streaming costs, assuming no long-term data retention.",[31,601,603],{"id":602},"inter-broker-replication-costs","Inter-Broker Replication Costs",[47,605,606],{},"Inter-broker (cross-AZ) replication is a major cost driver for data streaming engines:",[337,608,609,612,615],{},[340,610,611],{},"RedPanda: Inter-broker replication is not free, leading to substantial costs when data must be copied across multiple availability zones.",[340,613,614],{},"AWS MSK: Inter-broker replication is free, but MSK instance pricing is significantly higher (e.g., $3.264 per hour for kafka.m7g.8xlarge vs $1.306 per hour for an on-demand m7g.8xlarge). The storage price of MSK is $0.10 per GB-month which is significantly higher than st1, which costs $0.045 per GB-month. Even though replication is free, client-to-broker traffic still incurs inter-AZ charges.",[340,616,617],{},"Ursa: No inter-broker replication costs due to its leaderless architecture, eliminating inter-zone replication costs entirely.",[31,619,621],{"id":620},"zone-affinity-reducing-inter-az-costs","Zone Affinity: Reducing Inter-AZ Costs",[47,623,624],{},"We evaluated zone affinity mechanisms to further reduce inter-AZ data transfer costs.",[47,626,627],{},"Consumers:",[337,629,630],{},[340,631,632],{},"Follower fetch is enabled across all tests, ensuring consumers fetch data from replicas in their local AZ—eliminating inter-zone consumer traffic except for metadata lookups",[47,634,635],{},"Producers:",[337,637,638,647,656],{},[340,639,640,641,646],{},"Kafka protocol lacks an easy way to enforce producer AZ affinity (though ",[54,642,645],{"href":643,"rel":644},"https:\u002F\u002Fcwiki.apache.org\u002Fconfluence\u002Fdisplay\u002FKAFKA\u002FKIP-1123:+Rack-aware+partitioning+for+Kafka+Producer",[263],"KIP-1123"," aims to address this). And it only works with the default partitioner (i.e., when no record partition or record key is specified).",[340,648,649,650,655],{},"Redpanda recently introduced ",[54,651,654],{"href":652,"rel":653},"https:\u002F\u002Fdocs.redpanda.com\u002Fredpanda-cloud\u002Fdevelop\u002Fproduce-data\u002Fleader-pinning\u002F",[263],"leader pinning",", but this only benefits setups where producers are confined to a single AZ—not applicable to our multi-AZ benchmark.",[340,657,658,659,664],{},"Ursa is the only system in this test with ",[54,660,663],{"href":661,"rel":662},"https:\u002F\u002Fdocs.streamnative.io\u002Fdocs\u002Fconfig-kafka-client#eliminate-cross-az-networking-traffic",[263],"built-in zone affinity for both producers and consumers",". It achieves this by embedding producer AZ information in client.id, allowing metadata lookups to route clients to local-AZ brokers, eliminating inter-AZ producer traffic.",[31,666,668],{"id":667},"cost-comparison-results","Cost Comparison Results",[47,670,335],{},[337,672,673,675],{},[340,674,342],{},[340,676,345],{},[47,678,679],{},"Ursa’s leaderless architecture, zone affinity, and native cloud storage integration deliver unparalleled cost efficiency, making it the most cost-effective choice for high-throughput data streaming workloads.",[47,681,682],{},[349,683],{"alt":17,"src":684},"\u002Fimgs\u002Fblogs\u002F679c72208198ca36a352f228_AD_4nXeeZuM8T-xBlD4Vf3j67K618n08qh8wIDLLtiLJG0ssA1Wj1V26u7wIDTX9sqLrtw8mB2c299dwzarGen62CG0Vh7nWstn5qbPGFcBaKJYEepTsLr5fHWv1U8uqbg8Y0UOK6fJ7.png",[47,686,687],{},[349,688],{"alt":17,"src":689},"\u002Fimgs\u002Fblogs\u002F679c625978031f40229de484_AD_4nXdLkLLJ30KKr-_A_rN1j8akVwBYacAWIPzWHoOReJF421890kfByZoQQxkLczihVSmiw5Q9J51-V9I2SEKITbwsYnANDDTlAVL5nQ_jfaHNTe9VEWhSoa7DZooCnilDYL6l6msmJg.png",[47,691,692],{},"The detailed infrastructure cost calculations for each data streaming engine are listed below:",[31,694,696],{"id":695},"streamnative-ursa","StreamNative - Ursa",[337,698,699,702,705,708],{},[340,700,701],{},"Server EC2 costs: 9 * $1.536\u002Fhr = $14",[340,703,704],{},"Client EC2 costs: 9 * $1.536\u002Fhr =$14",[340,706,707],{},"S3 write requests costs: 1350 r\u002Fs * $0.005\u002F1000r * 3600s = $24",[340,709,710],{},"S3 read requests costs: 1350 r\u002Fs * $0.0004\u002F1000r * 3600s = $2",[31,712,714],{"id":713},"aws-msk","AWS MSK",[337,716,717,720,723],{},[340,718,719],{},"Server EC2 costs: 15 * $3.264\u002Fhr = $49",[340,721,722],{},"Client side EC2 costs: 9 * $1.536\u002Fhr =$14",[340,724,725],{},"Interzone traffic - producer to broker: 5GB\u002Fs * ⅔ * $0.02\u002FG(in+out) * 3600 = $240",[31,727,729],{"id":728},"redpanda","RedPanda",[337,731,732,734,736,739,742],{},[340,733,701],{},[340,735,704],{},[340,737,738],{},"Interzone traffic - producer to broker: 5GB\u002Fs * ⅔ * $0.02\u002FGB(in+out) * 3600 = $240",[340,740,741],{},"Interzone traffic - replication: 10GB\u002Fs * $0.02\u002FGB(in+out) * 3600 = $720",[340,743,744],{},"Interzone traffic - broker to consumer: $0 (fetch from local zone)",[47,746,747,748,753],{},"Please note that we were unable to test ",[54,749,752],{"href":750,"rel":751},"https:\u002F\u002Fwww.redpanda.com\u002Fblog\u002Fcloud-topics-streaming-data-object-storage",[263],"Redpanda with Cloud Topics",", as it remains an announced but unreleased feature and is not yet available for evaluation. Based on the limited information available, while Cloud Topics may help optimize inter-zone data replication costs, producers still need to traverse inter-availability zones to connect to the topic partition owners and incur inter-zone traffic costs of up to $240 per hour.",[337,755,756,762],{},[340,757,758,761],{},[54,759,645],{"href":643,"rel":760},[263]," (when implemented) will help mitigate producer-to-broker inter-zone traffic, but it is not yet available. And it only works with the default partitioner (no record partition or key is specified).",[340,763,764],{},"Redpanda’s leader pinning helps only when all producers for the pinned topic are confined to a single AZ. In multi-AZ environments (like our benchmark), inter-zone producer traffic remains unavoidable.",[47,766,767],{},"Additionally, Redpanda’s Cloud Topics architecture is not documented publicly. Their blog mentions \"leader placement rules to optimize produce latency and ingress cost,\" but it is unclear whether this represents a shift away from a leader-based architecture or if it uses techniques similar to Ursa’s zone-aware approach.",[47,769,770],{},"We may revisit this comparison as more details become available.",[39,772,774],{"id":773},"comparing-total-cost-of-ownership","Comparing Total Cost of Ownership",[47,776,777],{},"As highlighted earlier, with a BYOC Ursa setup, you can achieve 5 GB\u002Fs throughput at just 5% of the infrastructure cost of a traditional leader-based data streaming engine, such as Kafka or RedPanda, while managing the infrastructure yourself. This significant cost reduction is enabled by Ursa’s leaderless architecture and lakehouse-native storage design, which eliminate overhead costs such as inter-zone traffic and leader-based data replication. By leveraging a lakehouse-native, leaderless architecture, Ursa reduces resource requirements, enabling you to handle high data throughput efficiently and at a fraction of the cost of RedPanda.",[47,779,780],{},"Now, let’s examine the total cost comparison, evaluating Ursa alongside other vendors, including those that have adopted a leaderless architecture (e.g., Confluent WarpStream). This comparison is based on a 5GB\u002Fs workload with a 7-day retention period, factoring in both storage cost and vendor costs Here are the key findings:",[337,782,783,786,789],{},[340,784,785],{},"Ursa ($164,353\u002Fmonth) is: 50% cheaper than Confluent WarpStream ($337,068\u002Fmonth)",[340,787,788],{},"85% cheaper than AWS MSK ($1,115,251\u002Fmonth)",[340,790,791],{},"86% cheaper than Redpanda ($1,202,853\u002Fmonth)",[47,793,794],{},"In addition to Ursa’s architectural advantages—eliminating most inter-AZ traffic and leveraging lakehouse storage for cost-effective data retention—it also adopts a more fair and cost-efficient pricing model: Elastic Throughput-based pricing. This approach aligns costs with actual usage, avoiding unnecessary overhead.",[47,796,797],{},"Unlike WarpStream, which charges for both storage and throughput, Ursa ensures that customers only pay for the throughput they actively use. Ursa’s pricing is based on compressed data sent by clients, meaning the more data compressed on the client side, the lower the cost. In contrast, WarpStream prices are based on uncompressed data, unfairly inflating expenses and failing to incentivize customers to optimize their client applications.",[47,799,800],{},"This distinction is crucial, as compressed data reduces both storage and network costs, making Ursa’s pricing model not only more cost-effective but also more transparent and predictable.",[47,802,803],{},[349,804],{"alt":17,"src":805},"\u002Fimgs\u002Fblogs\u002F679c602d194800c9206d9d58_AD_4nXcFlf755xgyz7htxhMhBV5fGrsxy642mQNodt61DTok_z1dwkw5A6lkO5hatXVneCaB0anbZPAyvLI3MlIMuQEYLEACHHvQMOr5UfaB37dfzkdqewDEvcT-20VGd_zzvJsuA00zGA.png",[47,807,808],{},[349,809],{"alt":17,"src":810},"\u002Fimgs\u002Fblogs\u002F679c62594e9c2e629fae73aa_AD_4nXeU6cOgItnjLsEZCOf13TEvMY_SHWWIxYP2OYUj-B1GUPyWO78OG08K_v03hwYSVcg06f9dqDiGmdwy76vynjmiDGL5bluZ5_XF4nSU_r59oOZdfViXndXt6s11vVOY7qwfZN8v.png",[31,812,814],{"id":813},"cost-breakdown","Cost Breakdown",[816,817,818],"h4",{"id":695},"StreamNative – Ursa",[337,820,821,824,827,830,833],{},[340,822,823],{},"EC2 (Server): 9 × $1.536\u002Fhr × 24 hr × 30 days = $9,953.28",[340,825,826],{},"S3 Write Requests: 1,350 r\u002Fs × $0.005\u002F1,000 r × 3,600 s × 24 hr × 30 days = $17,496",[340,828,829],{},"S3 Read Requests: 1,350 r\u002Fs × $0.0004\u002F1,000 r × 3,600 s × 24 hr × 30 days = $1,400",[340,831,832],{},"S3 Storage Costs: 5 GB\u002Fs × $0.021\u002FGB × 3,600 s × 24 hr × 7 days = $63,504",[340,834,835],{},"Vendor Cost: 200 ETU × $0.50\u002Fhr × 24 hr × 30 days = $72,000",[816,837,839],{"id":838},"warpstream","WarpStream",[337,841,842,845],{},[340,843,844],{},"Based on WarpStream’s pricing calculator (as of January 29, 2025), we assume a 4:1 client data compression ratio, meaning 20 GB\u002Fs of uncompressed data translates to 5 GB\u002Fs of compressed data.",[340,846,847,848,853],{},"It's important to note that WarpStream’s pricing structure has fluctuated frequently throughout January. We observed the cost reported by their calculator changing from $409,644 per month to $337,068 per month. This variability has been previously highlighted in the blog post “",[54,849,852],{"href":850,"rel":851},"https:\u002F\u002Fbigdata.2minutestreaming.com\u002Fp\u002Fthe-brutal-truth-about-apache-kafka-cost-calculators",[263],"The Brutal Truth About Kafka Cost Calculators","”. To ensure transparency, we have documented the pricing as of January 29, 2025.",[47,855,856],{},[349,857],{"alt":17,"src":858},"\u002Fimgs\u002Fblogs\u002F679c602e42713e0028e9af5e_AD_4nXcu5_VWTLu9jRYs6zX1MBAOtLQEo5gyfNSWPcbpnQHXTa8qNCFAXezRR2E8daygzYTTwd4dhJjaLaLM8C6y_3OGbu2NS7pdvEv3a8-ptNKOg7AeKnYqPQCAYvQ5EuxzuI3JYIvY.png",[816,860,862],{"id":861},"msk","MSK",[337,864,865,868,871],{},[340,866,867],{},"EC2 (Server): 15 * $3.264\u002Fhr × 24 hr × 30 days = $35,251",[340,869,870],{},"Interzone Traffic (Client-Server): 5 GB\u002Fs × ⅔ × $0.02\u002FGB (in+out) × 3,600 s × 24 hr × 30 days = $172,800",[340,872,873],{},"Storage: 5 GB\u002Fs × $0.1\u002FGB-month × 3,600 s × 24 hr × 7 days * 3 replicas = $907,200",[816,875,729],{"id":876},"redpanda-1",[337,878,879,882,884,887,890],{},[340,880,881],{},"EC2 (Server): 9 × $1.536\u002Fhr × 24 hr × 30 days = $9953",[340,883,870],{},[340,885,886],{},"Interzone Traffic (Replication): 5 GB\u002Fs × 2 × $0.02\u002FGB (in+out) × 3,600 s × 24 hr × 30 days = $518,400",[340,888,889],{},"Storage: 5 GB\u002Fs × $0.045\u002FGB-month(st1) × 3,600 s × 24 hr × 7 days * 3 replicas = $408,240",[340,891,892],{},"Vendor Cost: $93,333 per month (based on limited information. See additional notes below).",[816,894,896],{"id":895},"additional-notes","Additional Notes",[337,898,899],{},[340,900,901,902,907],{},"Redpanda does not publicly disclose its BYOC pricing, making it difficult to accurately assess its total costs. We refer to information from the whitepaper “",[54,903,906],{"href":904,"rel":905},"https:\u002F\u002Fwww.redpanda.com\u002Fresources\u002Fredpanda-vs-confluent-performance-tco-benchmark-report#form",[263],"Redpanda vs. Confluent: A Performance and TCO Benchmark Report by McKnight Consulting Group.","” for estimation purposes. Based on the Tier-8 pricing model in the whitepaper,  the estimated cost to support a 5GB\u002Fs workload would be $1.12 million per year ($93,333 per month). However, since this calculation is based on an estimation, we will revisit and refine the cost assessment once Redpanda publishes its BYOC pricing.",[47,909,910],{},[349,911],{"alt":17,"src":912},"\u002Fimgs\u002Fblogs\u002F679c602dc8a9859eed89a0ef_AD_4nXdbcO8vsNNPy4GtkNLlmNKf22fjxRvzLzH7CtOna1L08sTbvnZx3HhufeFqc1w4K2gEF7lxO2IR5supotxebAiGnA07Qa8Yr3Rd1pVK2LYKK4WurlJGwgdwwucZIFoF-N_2oBjY.png",[47,914,915],{},[349,916],{"alt":17,"src":917},"\u002Fimgs\u002Fblogs\u002F679c602d6bc1c2287e012540_AD_4nXfcHZnLfjbjIr3ZAgoQXT9dwP3aQCOQPmGZZJUtpNZSwE6qY6M3yehIaBxCwxEIeu5PVdUPY0zhyjnow26YfgjdYgSG4GnV9ibxu0YWTIpwng6z_F6FUGJMpERMKtpsFESzXSN_Sw.png",[337,919,920,923],{},[340,921,922],{},"When estimating the storage costs for Kafka and Redpanda, we assume the use of HDD storage at $0.045\u002FGB, based on the premise that both systems can fully utilize disk bandwidth without incurring the higher costs associated with GP2 or GP3 volumes. However, in practice, many users opt for GP2 or GP3, significantly increasing the total storage cost for Kafka and Redpanda.",[340,924,925],{},"Unlike disk-based solutions, S3 storage does not require capacity preallocation—Ursa only incurs costs for the actual data stored. This contrasts with Kafka and Redpanda, where preallocating storage can drive up expenses. As a result, the real-world storage costs for Kafka and Redpanda are often 50% higher than the estimates above.",[39,927,929],{"id":928},"conclusion","Conclusion",[47,931,932],{},"Ursa represents a transformative shift in streaming data infrastructure, offering cost efficiency, scalability, and flexibility without compromising durability or reliability. By leveraging a leaderless architecture and eliminating inter-zone data replication, Ursa reduces total cost of ownership by over 90% compared to traditional leader-based streaming engines like Kafka and Redpanda. Its direct integration with cloud storage and scalable metadata & index management via Oxia ensure high availability and simplified infrastructure management.",[31,934,936],{"id":935},"balancing-latency-and-cost","Balancing Latency and Cost",[47,938,939,943],{},[54,940,942],{"href":941},"\u002Fblog\u002Fcap-theorem-for-data-streaming","Ursa trades off slightly higher latency for ultra low cost",", making it an ideal choice for the majority of streaming workloads, especially those that prioritize throughput and cost savings over ultra-low latency. Meanwhile, StreamNative’s BookKeeper-based engine remains the preferred solution for real-time, latency-sensitive applications. By combining these two approaches, StreamNative empowers customers with the flexibility to choose the right engine for their specific needs—whether it's maximizing cost savings or achieving ultra low-latency real-time performance.",[31,945,947],{"id":946},"the-future-of-streaming-infrastructure","The Future of Streaming Infrastructure",[47,949,950],{},"In an era where data fuels AI, analytics, and real-time decision-making, managing infrastructure costs is critical to sustaining innovation. Ursa is not just a cost-cutting alternative—it is a forward-thinking, lakehouse-native platform that redefines how modern data streaming infrastructure should be built and operated.",[47,952,953,954,959],{},"Whether your priority is reducing costs, improving flexibility, or ingesting massive data into lakehouses, Ursa delivers a future-proof solution for the evolving demands of real-time data streaming. ",[54,955,958],{"href":956,"rel":957},"https:\u002F\u002Fconsole.streamnative.cloud\u002F",[263],"Get started"," with StreamNative Ursa today!",[961,962,964],"h1",{"id":963},"references","References",[47,966,967,970,971],{},[968,969,428],"span",{}," ",[54,972,973],{"href":973},"\u002Fblog\u002Fintroducing-oxia-scalable-metadata-and-coordination",[47,975,976,970,978],{},[968,977,377],{},[54,979,376],{"href":376},[47,981,982,970,985],{},[968,983,984],{},"StreamNative pricing",[54,986,987],{"href":987,"rel":988},"https:\u002F\u002Fdocs.streamnative.io\u002Fdocs\u002Fbilling-overview",[263],[47,990,991,970,994],{},[968,992,993],{},"WarpStream pricing",[54,995,996],{"href":996,"rel":997},"https:\u002F\u002Fwww.warpstream.com\u002Fpricing#pricingfaqs",[263],[47,999,1000,970,1003],{},[968,1001,1002],{},"AWS S3 pricing",[54,1004,1005],{"href":1005,"rel":1006},"https:\u002F\u002Faws.amazon.com\u002Fs3\u002Fpricing\u002F",[263],[47,1008,1009,970,1012],{},[968,1010,1011],{},"AWS EBS pricing",[54,1013,1014],{"href":1014,"rel":1015},"https:\u002F\u002Faws.amazon.com\u002Febs\u002Fpricing\u002F",[263],[47,1017,1018,970,1021],{},[968,1019,1020],{},"AWS MSK pricing",[54,1022,1023],{"href":1023,"rel":1024},"https:\u002F\u002Faws.amazon.com\u002Fmsk\u002Fpricing\u002F",[263],[47,1026,1027,970,1030],{},[968,1028,1029],{},"The Brutal Truth about Kafka Cost Calculators",[54,1031,850],{"href":850,"rel":1032},[263],[47,1034,1035,970,1038],{},[968,1036,1037],{},"Redpanda vs. Confluent: A Performance and TCO Benchmark Report by McKnight Consulting Group",[54,1039,904],{"href":904,"rel":1040},[263],{"title":17,"searchDepth":18,"depth":18,"links":1042},[1043,1044,1045,1050,1054,1055,1064,1067],{"id":331,"depth":18,"text":332},{"id":370,"depth":18,"text":371},{"id":395,"depth":18,"text":396,"children":1046},[1047,1048,1049],{"id":407,"depth":278,"text":408},{"id":432,"depth":278,"text":433},{"id":453,"depth":278,"text":454},{"id":477,"depth":18,"text":478,"children":1051},[1052,1053],{"id":481,"depth":278,"text":482},{"id":496,"depth":278,"text":497},{"id":537,"depth":18,"text":538},{"id":549,"depth":18,"text":550,"children":1056},[1057,1058,1059,1060,1061,1062,1063],{"id":556,"depth":278,"text":557},{"id":602,"depth":278,"text":603},{"id":620,"depth":278,"text":621},{"id":667,"depth":278,"text":668},{"id":695,"depth":278,"text":696},{"id":713,"depth":278,"text":714},{"id":728,"depth":278,"text":729},{"id":773,"depth":18,"text":774,"children":1065},[1066],{"id":813,"depth":278,"text":814},{"id":928,"depth":18,"text":929,"children":1068},[1069,1070],{"id":935,"depth":278,"text":936},{"id":946,"depth":278,"text":947},"StreamNative Cloud","2025-01-31","Discover how Ursa achieves 5GB\u002Fs Kafka workloads at just 5% of the cost of traditional streaming engines like Redpanda and AWS MSK. See our benchmark results comparing infrastructure costs, total cost of ownership (TCO), and performance across leading Kafka vendors.","\u002Fimgs\u002Fblogs\u002F679c6593d25099b1cdcec4ca_image-31.png",{},"\u002Fblog\u002Fhow-we-run-a-5-gb-s-kafka-workload-for-just-50-per-hour","30 min",{"title":306,"description":1073},"blog\u002Fhow-we-run-a-5-gb-s-kafka-workload-for-just-50-per-hour",[1081,1082,301],"TCO","Apache Kafka","A0o_2xdJiLI6rf6xj4RKsxJNo_A6QN2fYzCp6gaLrFw",{"id":1085,"title":1086,"authors":1087,"body":1090,"category":1604,"createdAt":10,"date":1605,"description":1606,"extension":8,"featured":292,"image":1607,"isDraft":292,"link":10,"meta":1608,"navigation":7,"order":294,"path":1609,"readingTime":1610,"relatedResources":10,"seo":1611,"stem":1612,"tags":1613,"__hash__":1615},"blogs\u002Fblog\u002Fhow-pulsars-architecture-delivers-better-performance-than-kafka.md","How Pulsar’s architecture delivers better performance than Kafka",[1088,1089],"Yiming Zang","Julien Jakubowski",{"type":14,"value":1091,"toc":1582},[1092,1096,1099,1102,1106,1109,1113,1116,1121,1124,1127,1130,1134,1137,1142,1145,1148,1151,1160,1163,1169,1172,1176,1179,1182,1196,1200,1203,1207,1210,1213,1224,1228,1231,1234,1239,1242,1246,1251,1254,1257,1260,1263,1266,1270,1273,1277,1280,1283,1287,1290,1293,1297,1300,1303,1306,1309,1320,1323,1331,1335,1338,1343,1346,1350,1353,1358,1361,1375,1378,1380,1384,1388,1391,1396,1399,1404,1407,1410,1419,1423,1426,1431,1434,1439,1442,1445,1448,1451,1462,1465,1470,1472,1476,1479,1482,1496,1499,1519,1522,1526,1529,1532,1535,1549,1554,1557,1560,1563,1566,1568,1570,1573,1576],[39,1093,1095],{"id":1094},"overview","Overview",[47,1097,1098],{},"In the realm of distributed messaging systems, Apache Kafka and Apache Pulsar stand out as popular choices for high-throughput, real-time data streaming. While both platforms excel in their respective capabilities, Pulsar has garnered attention for its remarkable speed.",[47,1100,1101],{},"This may be surprising given that the architecture of Pulsar is more sophisticated, notably involving the presence of an extra network hop between multiple layers. And yet, despite the presence of a network, Pulsar can outperform Kafka in terms of performance. This article explains how this is possible.",[31,1103,1105],{"id":1104},"understanding-the-architectural-differences","Understanding the Architectural Differences",[47,1107,1108],{},"To comprehend the disparities in performance, it is crucial to examine the architectural variances between Pulsar and Kafka.",[31,1110,1112],{"id":1111},"kafka-architecture","Kafka Architecture",[47,1114,1115],{},"Apache Kafka operates on a three-tier architecture with the presence of Application Clients, Kafka Brokers and ZooKeeper.",[47,1117,1118],{},[349,1119],{"alt":17,"src":1120},"\u002Fimgs\u002Fblogs\u002F64dcd185982c22533d1ddaab_8-H9QT1d7gJ0-21EhE9rrdXVy6HMCUOLlpu7GiT2aTpYt-MivAYRANWyatgLVyKwzxynNxdI_5Lo_55yQhyMZ7f2nBYiaA9rLqoukZdS6NblRpq4yEMrZdZ3cV54P4F861xQWQe2yuKJtyoglYnfo7E.png",[47,1122,1123],{},"Kafka Producer and Consumer clients connect directly to Kafka brokers for reads and writes. Zookeeper serves as the metadata layer providing partition ownership and leader election, which doesn’t directly serve in the critical read or write path.",[47,1125,1126],{},"Additionally, by utilizing Kraft in a Kafka cluster, the traditional need for Zookeeper nodes could be eliminated in Kafka, resulting in a two-tier architecture.",[47,1128,1129],{},"‍",[31,1131,1133],{"id":1132},"pulsar-architecture","Pulsar Architecture",[47,1135,1136],{},"In comparison, Apache Pulsar adopts a four-tier architecture comprising Clients, Pulsar Brokers, Apache Bookkeeper, and a configurable Metadata Store(e.g ZooKeeper, Etcd, RocksDB or Oxia). In some cases, we would even need a five-tier architecture with a Pulsar Proxy.",[47,1138,1139],{},[349,1140],{"alt":17,"src":1141},"\u002Fimgs\u002Fblogs\u002F64dcd1851093465ed076eb03_x46m91bNCzjfpvlrhjCSLS8SU2Wzxrxb-YTh8lfggRhxvT9UlflVgS8vLQ_4EUqamP4j-7K6NK9v4zdGGK706_Q0shmd5tWiaJAY1xM4h56E_ZBIWCXB88Vvc8yGDVf4B9j6Wp7fm78jXd5NoPEf7Es.png",[47,1143,1144],{},"Pulsar Client can connect directly to Brokers to produce and consume messages as well as for topic lookup.",[47,1146,1147],{},"Pulsar proxy is an optional gateway component, which can be used when direct connections between Clients and Pulsar brokers are either infeasible or undesirable. For example, StreamNative Cloud adopts Istio in place of Pulsar Proxy to achieve high availability and performance.",[47,1149,1150],{},"Pulsar Brokers is a stateless component which serves mainly for two purposes:",[1152,1153,1154,1157],"ol",{},[340,1155,1156],{},"Topic lookup, which tells you which topic partition is owned by which broker",[340,1158,1159],{},"Dispatcher, which transfers data and dispatch them via managed ledger by talking to Bookkeeper",[47,1161,1162],{},"Bookkeeper is simply the storage layer, similar to Kafka brokers in the sense where it persists all the data and serves reads and writes.",[47,1164,1165,1166,1168],{},"Pulsar supports configurable Metadata Store (e.g ZooKeeper, Etcd, RocksDB, Oxia). Historically, Apache ZooKeeper has been the most popular primary Pulsar metadata store. StreamNative recently invented ",[54,1167,428],{"href":973},", a  better scalable metadata store. This metadata store is very critical to Pulsar since it is being used for coordination, and storing key metadata information such as topic partition ownership, Bookkeeper ledger metadata, etc.",[47,1170,1171],{},"When comparing the architectures of Pulsar and Kafka, it's noticeable that Pulsar assigns more granular roles to its components. However, intriguingly, Pulsar still manages to outperform Kafka in most scenarios. In the upcoming sections, we'll explore the factors contributing to Pulsar's impressive performance despite this architectural difference.",[39,1173,1175],{"id":1174},"kafka-lack-of-isolation","Kafka - Lack of Isolation",[47,1177,1178],{},"One prominent factor contributing to Kafka's relatively slower performance lies in its design.",[47,1180,1181],{},"Kafka does not inherently provide good IO isolation, leading to substantial interference between read and write traffic.",[337,1183,1184,1187,1190,1193],{},[340,1185,1186],{},"Read and writes can impact each other: Kafka brokers do not have separate dedicated IO threads exclusively for read and write requests. Besides, high write throughput can lead to increased disk and I\u002FO pressure on Kafka brokers which can affect read performance, and vice versa, high read rates from consumers can potentially impact write performance as well.",[340,1188,1189],{},"Lack of Isolation among different topics or partitions: There is no hard separation or resource isolation between topics or partitions, and they can all compete with each other for disk and network IO. This means one hot topic\u002Fpartition can impact its local neighbors' performance a lot in a persistent manner, which makes Kafka difficult to be maintained in a multi-tenant environment",[340,1191,1192],{},"Lack of isolation between disks, cpu and network resources: Kafka brokers act as both a serving layer and the persistence storage layer. You can not scale them out independently, which means you will always hit the bottleneck for one of the resources first, either the disk first or the cpu\u002Fnetwork first. For on-prem users, it’s fairly difficult to tune resource allocation to perfectly fit the use case or traffic pattern. For cloud users, unfortunately, Kafka itself is not very cloud friendly, e.g. running Kafka on K8S is still a challenge in terms of operation.",[340,1194,1195],{},"Tightly coupled partitioning model: Kafka partitions need to be scaled accordingly for the consumer jobs based on the data processing speed.  However, having too many partitions at the same time will hurt batch efficiency as well as compression rate. So it is sometimes hard to tune the number of partitions for a kafka topic.",[39,1197,1199],{"id":1198},"pulsar-better-isolation","Pulsar - Better Isolation",[47,1201,1202],{},"Apache Pulsar has been designed with a focus on achieving strong IO isolation, which is one of its key architectural strengths.",[31,1204,1206],{"id":1205},"separating-compute-from-storage","Separating Compute from Storage",[47,1208,1209],{},"Pulsar’s architecture allows for independent scaling of Brokers for compute and network, and scaling Bookies for disk space or IO, which provides a much better isolation and decoupling between disk and network throughput limitation.",[47,1211,1212],{},"This would be beneficial for certain scenarios, such as:",[337,1214,1215,1218,1221],{},[340,1216,1217],{},"Read heavy or high fan-out, but write throughput is low: You can independently scale out Brokers to handle more reads",[340,1219,1220],{},"Extremely short or long retention: You can dynamically scale in or out bookies based on disk usage you need",[340,1222,1223],{},"Write heavy or high fan-in, but reads are small: You can independently scale out bookies to handle more write throughput",[31,1225,1227],{"id":1226},"segmented-storage","Segmented Storage",[47,1229,1230],{},"Writes to a topic partition are split into segments, which are then stripped across multiple bookie nodes, instead of a single bookie.",[47,1232,1233],{},"This provides a better isolation between topics and partitions so that one hot topic or partition will not keep impacting other topics living on the same node.",[47,1235,1236],{},[349,1237],{"alt":17,"src":1238},"\u002Fimgs\u002Fblogs\u002F64dcd189f25c6bb275c31711_LTbd20CEAXx4_HkXBkvXeLEk8oVeg6fxDapyAKd2HTCgzv8JQD5yu1G1WkPC-0NoWExlATjMODHsU39VBOkDH54bCA0jYCZYxZ7UQaIH2YkzBJWC7ilGX0KdYdssILlnnCHIymEHjw4TVbwJsoGbc88.png",[47,1240,1241],{},"To be more specific, the data of any Pulsar partition will be spread across the whole bookkeeper cluster, unlike Kafka, all of the data of a single partition always stays on the same set of brokers. The benefit of this is that, if a single partition becomes really hot or overloaded, it’s not an issue for Pulsar because load will be spread evenly to all Bookies, but it’s a big problem for Kafka because it will overload the three Kafka brokers which own that partition, and thus cause issues for other topics owned by the same set of brokers.",[31,1243,1245],{"id":1244},"bookkeeper-io-isolation","Bookkeeper IO Isolation",[47,1247,1248],{},[349,1249],{"alt":17,"src":1250},"\u002Fimgs\u002Fblogs\u002F64dcd186c6b49b490adff562_3p_DpJnB_1M_Y-4nYX_Grzk7YdwAK2kno-icx6jNCqcIhBvuV_amY3rS1A8WEmT-6hgxr-Ij2EiNUUvDkSTuoEIVQ1OypZAcxQjBH9W0cwo5ZrpQVNAzNodwDs8b6b-_q_EJez8OH1rGbkcx1qSjh7s.png",[47,1252,1253],{},"Write Ahead Log (Journal)",[47,1255,1256],{},"BookKeeper uses a write-ahead log mechanism for durability. Data is first written to the Journal in sequential order and append-only manner before being persisted to the main ledger disk.",[47,1258,1259],{},"Storage Separation",[47,1261,1262],{},"BookKeeper separates the storage device used for the journal from the main ledger storage and the journal is usually stored on faster and more durable storage (e.g., SSDs) to handle the write-intensive workload effectively.",[47,1264,1265],{},"Tailing reads are always served from the memTable, and only catch up reads will be served from the Ledger Disk and Index Disk. As a result, heavy reads in Bookkeeper will not impact incoming write performance because they are served from different physical disks, and have pretty good isolation.",[31,1267,1269],{"id":1268},"summary","Summary",[47,1271,1272],{},"In summary, even though Kafka requires one or two less network hops than Pulsar, the latency overhead brought by read-write interference could potentially be a few magnitudes higher than network hop latency. Therefore, Kafka's performance can be influenced a lot by inefficiencies in other areas rather than the number of network hops.",[39,1274,1276],{"id":1275},"network-hops","Network Hops",[47,1278,1279],{},"Compared to disk performance, network latency is always a bigger concern since the network can become unreliable and thus network latency can go extremely high. A frequent concern raised by Pulsar users revolves around the network could potentially act as a bottleneck in Pulsar's multi-tier architecture.",[47,1281,1282],{},"Now let’s deep dive into this topic and see whether network hops are a real concern for Pulsar or not",[31,1284,1286],{"id":1285},"network-hop-latency-can-be-very-low","Network Hop Latency can be very Low",[47,1288,1289],{},"Network hop latency can be optimized and tuned to a very low level if machines are close to each other or sharing the same network. For example, servers in the same datacenter or region can achieve a ping latency with less than 1ms. EC2 instances on AWS within the same region typically have a ping latency with less than 1ms as well even for hosts in different AZs, and similar for GCP.",[47,1291,1292],{},"With this data in mind, as long as Pulsar components are all deployed within the same region, one or two extra hops would just mean a few extra milliseconds of latency overhead, which can be almost ignorable.",[31,1294,1296],{"id":1295},"write-path","Write Path",[47,1298,1299],{},"Just with the fact that Pulsar has a more complicated multi-layer architecture doesn’t mean it has more hops in its read and write critical paths. The actual end to end latency for publishing and consuming depends mainly on how it works under the hood.",[47,1301,1302],{},"So let’s dive more into how Kafka and Pulsar write paths look like.",[47,1304,1305],{},"Assuming the below replication settings being used:",[47,1307,1308],{},"Kafka Broker:",[337,1310,1311,1314,1317],{},[340,1312,1313],{},"ack=all",[340,1315,1316],{},"replication.factor=3",[340,1318,1319],{},"min.insync.replicas=2",[47,1321,1322],{},"Pulsar:",[337,1324,1325,1328],{},[340,1326,1327],{},"write quorum size=3",[340,1329,1330],{},"ack quorum size=2",[816,1332,1334],{"id":1333},"kafka-leader-follower-replication","Kafka: Leader Follower Replication",[47,1336,1337],{},"Kafka relies on a poll model, where write requests are sent to the leader broker first. The leader broker then has to wait for ALL of its ISR followers to fetch and replicate data, and acknowledge them back. This is more error prone because one slow follower Broker in the ISR can cause write operations to become significantly slower or even time out, impacting overall performance. And if the Leader Broker is experiencing some slowness, writes will also be impacted.",[47,1339,1340],{},[349,1341],{"alt":17,"src":1342},"\u002Fimgs\u002Fblogs\u002F64dd0b67ab10f4ad122b02a2_KZRZqqxFl1ddQHlqGj9Bm8zCtqrCEejfPhlfwtgLrCfnbtKX0qbSoBJe06GFbm08u0amn32RHCjLyYsykjhso_02lBafqspiePHr7KBKxvprTfRHVS4rAodVwPtsvRvrA9XYuNCEt33J9XtFmh_Xlv8.png",[47,1344,1345],{},"Of course, we can choose to reduce replica.lag.time.max.ms config so that if the follower broker becomes too slow, it will drop out of ISR and then we will be able to satisfy the write requests faster with only 2 brokers. However it also means that you will observe under replicated partitions and constant ISR expand\u002Fshrink operations, which impacts overall durability and reliability.",[816,1347,1349],{"id":1348},"pulsar-parallel-replication","Pulsar: Parallel Replication",[47,1351,1352],{},"In contrast, when writing to BookKeeper, Pulsar leverages a parallel replication strategy, where it sends the writes to all 3 bookies at the same time waiting for 2 acknowledgments.",[47,1354,1355],{},[349,1356],{"alt":17,"src":1357},"\u002Fimgs\u002Fblogs\u002F64dd0bd577cacd54094f9d86_FezkGPM3pEkWt-5ODFC7ybSL9p3Rp1MvR74OVbTQ-491oAidHEo6XilxjzFVE5pi8SCYvkfaOgZRFORSromTeORfgVelGHMRBh-XnR8QvKZBhf59fFYJjzVsxhbXJ4XuEeifXmchC658q5wslT1USjg.png",[47,1359,1360],{},"Since writes happen in parallel, with regards to write operations, both Pulsar and Kafka have to go through four hops end to end as shown in the above diagrams. Those hops on high level are:",[1152,1362,1363,1366,1369,1372],{},[340,1364,1365],{},"Client send a produce request to Kafka\u002FPulsar broker",[340,1367,1368],{},"Kafka\u002FPulsar Broker replicate data by talking to the other replicas",[340,1370,1371],{},"Kafka\u002FPulsar Broker get the acknowledgement response from other replicas",[340,1373,1374],{},"Kafka\u002FPulsar Broker send the write response back to the client",[47,1376,1377],{},"In conclusion, the major difference between Kafka and Pulsar for writes path is how the underline replication works, and the end to end number of hops are actually the same for both.",[47,1379,1129],{},[31,1381,1383],{"id":1382},"read-path","Read Path",[816,1385,1387],{"id":1386},"kafka-read-from-leader","Kafka: Read from Leader",[47,1389,1390],{},"Let’s first look at how Kafka fetch requests are being served.",[47,1392,1393],{},[349,1394],{"alt":17,"src":1395},"\u002Fimgs\u002Fblogs\u002F64dcd18595ea02055a076cc2_z1FJ1-iAvP5B5QKCmWxQuHdXMxlPRxIMaxLIeFW0oEuhTgWa_ati9f_eYn6kj7SazBEXNsaivh5Av9byzQi6noHPsXo1OGbjQAx7APw-Hj7Z5UlB4xayDTxkd6gFWuSMI8HmkWyPe2_NdR0RjYkRkEU.png",[47,1397,1398],{},"For fetch requests, the Kafka client will send it to the leader broker and just wait for the leader to send the response back, so there are only 2 hops which is extremely simple.",[47,1400,1401],{},[349,1402],{"alt":17,"src":1403},"\u002Fimgs\u002Fblogs\u002F64dcd18537fd18a6625a4b54_4zSCNuAnTw0UGZiUx1y42GAaHvZ0yHI2UbZJbVbFfXpL_PhOUyKE_gbJcG3ykXs7quv8gZ85FGGZKvnVKKq-1Cil7IYhucmf3M4Sx6HP5paQPhiECig8patnJR-AuCPTTfUI5kM1nN_JGvFh5uj8pBg.png",[47,1405,1406],{},"If we look into what happens inside the leader broker, when the leader broker receives the FetchRequest, it will first try to fetch the batch of records from PageCache, and if it can’t find the data from PageCache, then it will start seeking the data from the disks.",[47,1408,1409],{},"This overall process looks simple, but one critical drawback for this approach is that it has a very tight dependency on the leader broker being healthy and responsive, not being overloaded or having some network issues.",[47,1411,1412,1413,1418],{},"As we know that leader election will only happen when the leader becomes totally unresponsive, but not being slow. Kafka by default doesn’t allow you to read from any follower replicas, so if the leader encounters any failures or becomes slow, it will slow down all the fetch requests by a lot and consumers will start falling behind. Although this can be improved by adopting a specific cross-region replica distribution model such as ",[54,1414,1417],{"href":1415,"rel":1416},"https:\u002F\u002Fcwiki.apache.org\u002Fconfluence\u002Fdisplay\u002FKAFKA\u002FKIP-392%3A+Allow+consumers+to+fetch+from+closest+replica",[263],"KIP-392",", it is not straightforward to configure and thus it is not very widely adopted by most Kafka users.",[816,1420,1422],{"id":1421},"pulsar-speculative-read","Pulsar: Speculative Read",[47,1424,1425],{},"Now let’s take a look at how Pulsar serves read requests.",[47,1427,1428],{},[349,1429],{"alt":17,"src":1430},"\u002Fimgs\u002Fblogs\u002F64dcd185982c22533d1ddb50_wVcJ9lOBUgWX8CLWxDYr0poJNti1sxXGIo0duwmFO6PAQtCqPe88-vluDJzL46CLkTzxs-31fkdBjp8sZoODwmDyaqVqVQzlO1WS8ZCtaiByLcRUmRWzHIYqd9xb8bhrNGzbnQPnrcTeVz3voWo1qQ4.png",[47,1432,1433],{},"Similar to the leader concept in Kafka, Pulsar also has an ownership concept for its brokers, so one of the Pulsar brokers will own the partition which the client is trying to consume from. Tailing reads would be served from the Pulsar Broker managed ledger cache, but in case of cache miss, Pulsar Broker will get the data from the Bookies via speculative read.",[47,1435,1436],{},[349,1437],{"alt":17,"src":1438},"\u002Fimgs\u002Fblogs\u002F64dcd185682ccbc3bd32a5ee_wu0lO6b2uiwXURQ0rwSMUNelBwqAF9KMK5plney6Uo-1rn_LiDlCASqpzQdrkbjfgZG6oyUFgl5kr1W-ptgYkD8pS3WJdcHJopc4CcPc2JesNNxiKZ0k0IRYwt5zzB4j80aYBBwrff6sWS0sE-cpGe0.png",[47,1440,1441],{},"As shown in the above graph, Bookies also have their own ledger cache to serve tailing reads, so most tailing read requests still won’t incur any disk read IOPs.",[47,1443,1444],{},"While Pulsar entails one potential additional network hop from the read perspective, it does not necessarily result in slower performance. This is because tailing read requests in Pulsar are primarily served from the Pulsar Broker managed ledger cache, rather than from Bookie storage layer. Thus, the extra network hop does not really impose a significant performance penalty for read operations.",[47,1446,1447],{},"Compared to Kafka, Pulsar's design has its own pros and cons.",[47,1449,1450],{},"Pros",[337,1452,1453,1456,1459],{},[340,1454,1455],{},"In most scenarios, serving tailing reads only need 2 hops for both Kafka and Pulsar",[340,1457,1458],{},"Kafka requests are sticky to the leader, whereas Pulsar can easily unload a partition and switch owner",[340,1460,1461],{},"Pulsar is resilient to single or multiple Bookie failures using speculative reads, where as Kafka can not serve reads when partition is offline",[47,1463,1464],{},"Cons",[337,1466,1467],{},[340,1468,1469],{},"When partition ownership change or catch up reads happens, Pulsar Broker will not have any cache in memory, so all reads will have to be served from Bookies, resulting in extra hops and network bandwidth usage",[47,1471,1129],{},[31,1473,1475],{"id":1474},"latency-sensitivity","Latency Sensitivity",[47,1477,1478],{},"Kafka is actually much more network latency sensitive than Pulsar, which means when there’s a network degradation, e.g. a bad\u002Fslow link, you would more often see latency impact in Kafka than in Pulsar for read and write.",[47,1480,1481],{},"One reason behind this is due to Kafka’s tightly coupled partitioning model. When a single broker becomes slow due to a degraded network, it will impact a lot of topic partitions across multiple brokers in the cluster.",[337,1483,1484,1487,1490,1493],{},[340,1485,1486],{},"As partition leader, its followers might have trouble fetching data for replication.",[340,1488,1489],{},"As a follower, it might have trouble to keep up with other leaders",[340,1491,1492],{},"Consumers can only fetch data from the partition leader, not the followers",[340,1494,1495],{},"Producers can only publish data to the partition leader, not the followers",[47,1497,1498],{},"Pulsar on the other hand is more resilient than Kafka in terms of network failures or degradation, this is benefited from several factors",[337,1500,1501,1504,1507,1510,1513,1516],{},[340,1502,1503],{},"Single slow bookie will have very minimum impact to the overall performance",[340,1505,1506],{},"Parallel Replication",[340,1508,1509],{},"Speculative Reads",[340,1511,1512],{},"Ensemble Change on write failures",[340,1514,1515],{},"Pulsar Broker and partitions are loosely coupled and topic partition can easily be unloaded",[340,1517,1518],{},"Auto Rebalance",[47,1520,1521],{},"Pulsar brokers are stateless, the broker load balancer can easily rebalance the topic ownerships based on the broker's dynamic load. This is also helpful when a broker is unavailable(network-partitioned, down), since orphan topics will be immediately reassigned to the available brokers.",[39,1523,1525],{"id":1524},"connection-limitations","Connection Limitations",[47,1527,1528],{},"Kafka suffers from fan-in connection limitations since, without a proxy layer, all connections are strictly routed to the partition leader. This limitation leads to significant garbage collection problems and restricts scalability to support a high number of producer connections.",[47,1530,1531],{},"To test out the connection limitation for Kafka brokers, we have done some specific benchmarks, in which we keep sending a fixed amount of write traffic, around 300k events\u002Fsecond with average event size of 500 bytes per broker, but through a different number of producer clients.",[47,1533,1534],{},"The Kafka benchmark is performed using the following bare metal setup:",[337,1536,1537,1540,1543,1546],{},[340,1538,1539],{},"CPU: Intel Xeon or AMD EPYC 2.1 GHz processor with 64 cores",[340,1541,1542],{},"Storage: Two 4TB RPM NVMe SSD drives (JBOD)",[340,1544,1545],{},"Memory: 512GB of RAM",[340,1547,1548],{},"Network: 100Gb Nic bandwidth",[47,1550,1551],{},[349,1552],{"alt":17,"src":1553},"\u002Fimgs\u002Fblogs\u002F64dcd185c0039ddb3b4a1c2f_d0ZdivC_Lfp3sTlg9Z_xoiC3cssf7xNtL6SA_3584-toJq6PjxQDj62iz3iRhU2oEMNsA4ztOibCpTQTwTd2onWzrTJqPrxMSg9YNAYYt9R2YvNdcg__wx2K1TP7YcHK8mRE9EDF2yQmUyw1HK7fp0M.png",[47,1555,1556],{},"The above Kafka benchmark result shows that as the total number of fan-in connections for each Kafka broker crossed over 100k, we started to see performance degradation. When the number of connections got close to 150k, we started to see more produce failures as well as high GC time (around 20 to 40 seconds)",[47,1558,1559],{},"Scaling the number of Kafka brokers just for connections is expensive, and it does not help a lot because when using round robin partitioning, each single client instance will still create a dedicated connection to every single broker. So adding more brokers doesn’t change the fact that the whole Kafka cluster can still only support up to around 120k-150k clients. The major reason why performance starts to degrade when connections are high is due to the GC overhead brought by extra connections.",[47,1561,1562],{},"In contrast, Pulsar overcomes this challenge by having stateless Pulsar Brokers as a proxy layer between client and storage, which can be independently scalable to handle fan-in connections.",[47,1564,1565],{},"With Pulsar brokers' stateless characteristic, adding new brokers is relatively cheap, and Pulsar Broker Load Balancer would seamlessly rebalance partitions in a bundle to the new brokers without the need to move data at the bookie layer. As a result, a Pulsar cluster can easily handle many more connections than a Kafka cluster with its more sophisticated architecture.",[47,1567,1129],{},[39,1569,929],{"id":928},[47,1571,1572],{},"Despite potential network bottlenecks in its architecture, Pulsar outperforms Kafka in terms of speed in a lot of scenarios. This advantage stems from efficient IO isolation, optimized read and write performance, expandable network bandwidth, and more scalable connection handling.",[47,1574,1575],{},"As both Pulsar and Kafka continue to evolve, it remains crucial for organizations to evaluate their strengths and weaknesses based on specific use cases and requirements. Understanding the nuances of their architectures empowers decision-makers to make informed choices when selecting the most suitable messaging system for their real-time data streaming needs.",[47,1577,1578],{},[1579,1580],"binding",{"value":1581},"cta-blog",{"title":17,"searchDepth":18,"depth":18,"links":1583},[1584,1589,1590,1596,1602,1603],{"id":1094,"depth":18,"text":1095,"children":1585},[1586,1587,1588],{"id":1104,"depth":278,"text":1105},{"id":1111,"depth":278,"text":1112},{"id":1132,"depth":278,"text":1133},{"id":1174,"depth":18,"text":1175},{"id":1198,"depth":18,"text":1199,"children":1591},[1592,1593,1594,1595],{"id":1205,"depth":278,"text":1206},{"id":1226,"depth":278,"text":1227},{"id":1244,"depth":278,"text":1245},{"id":1268,"depth":278,"text":1269},{"id":1275,"depth":18,"text":1276,"children":1597},[1598,1599,1600,1601],{"id":1285,"depth":278,"text":1286},{"id":1295,"depth":278,"text":1296},{"id":1382,"depth":278,"text":1383},{"id":1474,"depth":278,"text":1475},{"id":1524,"depth":18,"text":1525},{"id":928,"depth":18,"text":929},"Apache Pulsar","2023-08-16","Pulsar outperforms Kafka in terms of speed in a lot of scenarios. This advantage stems from efficient IO isolation, optimized read and write performance, expandable network bandwidth, and more scalable connection handling. ","\u002Fimgs\u002Fblogs\u002F64dd0f2ac53f5aded86ca0c4_Illustration-2.png",{},"\u002Fblog\u002Fhow-pulsars-architecture-delivers-better-performance-than-kafka","8 min",{"title":1086,"description":1606},"blog\u002Fhow-pulsars-architecture-delivers-better-performance-than-kafka",[1082,1614],"Multi-Tenancy","H0WumwRBLCe_azd5noGLXIVx9gqpse2cxYZAxYj-vrQ",[1617,1629],{"id":1618,"title":1088,"bioSummary":1619,"email":10,"extension":8,"image":1620,"linkedinUrl":10,"meta":1621,"position":1626,"stem":1627,"twitterUrl":10,"__hash__":1628},"authors\u002Fauthors\u002Fyiming-zang.md","Yiming Zang is a platform engineer at StreamNative and contributor to Apache Pulsar, KoP and Apache Bookkeeper","\u002Fimgs\u002Fauthors\u002Fyiming-zang.png",{"body":1622},{"type":14,"value":1623,"toc":1624},[],{"title":17,"searchDepth":18,"depth":18,"links":1625},[],"Platform Engineer at StreamNative","authors\u002Fyiming-zang","QDx-WX0GMBC-lb35sDW8B5Ak-KAvKVassjzPZOK1DxA",{"id":1630,"title":1089,"bioSummary":1631,"email":10,"extension":8,"image":1632,"linkedinUrl":1633,"meta":1634,"position":1639,"stem":1640,"twitterUrl":10,"__hash__":1641},"authors\u002Fauthors\u002Fjulien-jakubowski.md","Julien Jakubowski is a Developer Advocate at StreamNative with over 20+ years of experience as a developer, staff engineer, and consultant. He has built several complex systems with distributed, scalable, and event-driven architecture for various industrial sectors such as retail, finance, and manufacturing.","\u002Fimgs\u002Fauthors\u002Fjulien-jakubowski.jpeg","https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fjulienjakubowski\u002F",{"body":1635},{"type":14,"value":1636,"toc":1637},[],{"title":17,"searchDepth":18,"depth":18,"links":1638},[],"Developer Advocate, StreamNative","authors\u002Fjulien-jakubowski","JEar_AzlsJteW1n49p4JOxYH0QlEyYddzgohjZmFxxs",[1643,1650,1656],{"path":1644,"title":1645,"date":1646,"image":-1,"link":-1,"collection":1647,"resourceType":1648,"score":1649,"id":1644},"\u002Fblog\u002Femerging-patterns-in-data-streaming-insights-from-current-2023","Emerging Trends in Data Streaming: Insights from Current 2023","2023-10-02","blogs","Blog",0.667,{"path":1651,"title":1652,"date":1653,"image":1654,"link":-1,"collection":1647,"resourceType":1648,"score":1655,"id":1651},"\u002Fblog\u002Freducing-total-cost-of-ownership-tco-for-enterprise-data-streaming-and-messaging","Reducing Total Cost of Ownership (TCO) for Enterprise Data Streaming and Messaging","2023-06-30","\u002Fimgs\u002Fblogs\u002F649e8f101a32886a3ae38bc7_photo_38324684.jpg",0.5,{"path":1657,"title":1658,"date":1659,"image":1660,"link":-1,"collection":1661,"resourceType":1662,"score":1663,"id":1657},"\u002Fwebinars\u002Fdecoding-kafka-challenges-addressing-common-pain-points-in-kafka-deployments","Decoding Kafka Challenges: Addressing Common Pain Points in Kafka Deployments","2024-02-23","\u002Fimgs\u002Fwebinars\u002F65d8fed3d232db46863472d3_image%20(20).png","webinars","Webinar",0.44,1775494049583]