Firebolt

NYC here we come! Come see Connor Carreras and the gang at #awssummit booth 933 for a hands-on demo of Firebolt's: → Blazing fast query performance on large datasets → High concurrency support for interactive data apps → Elasticity for cost-effective scaling Plus: Test your skills with our 'How Big is Your Data Knowledge' booth challenge for a chance to win some ultra geeky prizes 😎 #AWS #DataWarehouse #DataAnalytics

Curious about the best data warehouses for low-latency analytics? Andrew Schetinin led his team at Agritask through an extensive PoC comparing PostgreSQL, Snowflake, Redshift, BigQuery, Clickhouse, TimescaleDB, Druid, Cassandra, and Firebolt. He shares the results along with his take on the strengths and weaknesses of each platform in his latest Medium post: https://lnkd.in/dv5KYvKg #DataWarehouse #LowLatency #DataEngineering

Tomorrow at #AWSSummit Tel Aviv: Visit us at booth B17 for a live demo of Firebolt’s latest features. Learn how Firebolt delivers insanely fast data analytics on AWS - while cutting your monthly bill. We'll demo: ✔️Millisecond query performance on terabytes of data ✔️Low latency with unparalleled price-for-performance ✔️High query throughput without performance degradation ✔️Simplicity and familiarity of SQL that follows PostgreSQL dialect closely #datawarehouse #dataanalytics #dataengineering

SQL’s slow. SQL’s stupid. We hear these claims every time a new shiny tool enters the market, only to realize five years later when the hype dies down that SQL is actually a good idea. In this super techie episode of the Data Engineering Show, Andy Pavlo, Associate Professor at Carnegie Mellon University, joins the bros to delve into database internals and optimization. Andy discusses leveraging ML for autonomous database optimization, using Postgres for practical applications, tuning production databases safely, and why SQL is here to stay. Tune in: https://lnkd.in/dTe5Qj2B #dataengineering #sql #database

As a solution architect for a data warehouse vendor, I've seen a ton of PoCs (both successful and less successful). Evaluating data warehouses can be complex, even for data engineers! Here are three tips to make your PoCs more effective: 1. Define Your Expectations. Know what you want to achieve before you start testing. Here are a few example success criteria from past Firebolt evaluations: ✅ Reduce query execution time to under 1.5 seconds for all test queries. ✅ Enable complex aggregations to run over 100GB of data in less than 1 second. ✅ Support 50 concurrent queries with a response time of less than 2 seconds. 2. Outline the Evaluation Process. Plan your evaluation process in advance to best simulate how your organization will interact with the platform. We typically cover: ✅ Ingesting initial data. ✅ Running test queries, including a mix of analytical queries with filters, aggregations, and joins. ✅ Optimizing performance using indexes. ✅ Testing integrations for ingestion and querying. Try your best to avoid “scope creep” as you test products. It’s easy to get distracted by cool features…but if those features aren’t required for your use case or representative of how your users will work with a data warehouse, those cool features are ultimately not valuable. 3. Benchmark During Ingestion and Querying. Gather benchmarks about your rates during these steps. This data can help you extrapolate your expected ongoing costs. One of the easiest ways to think about cost is "price per query", which you can only calculate if you've documented performance data. By following these steps, you'll get a solid feel for whether each data warehouse platform fits your organization's needs. #datawarehouse #dataengineering #dataanalytics 

I’m happy to share that I’m starting a new position as Head of Product Marketing (including core PMM function, technical content marketing, competitive intelligence, and developer relations) at Firebolt! I couldn't be more excited about what lies ahead as we build the next generation of cloud data warehouses for engineers, offering incredibly fast data analytics. --- After leading product marketing at Amazon Web Services (AWS), Microsoft, Neo4j, and Yellow.ai, Tara Shankar Jana "TJ" is joining the Firebolt A-Team 🔥 When he's not catching the latest highlights of Inter Miami soccer matches as a die-hard Messi fan (yes Mosha Pasumansky, we have another Messi fan joining Firebolt), he's busy sharpening his prompt engineering skills. Help us give him a warm welcome! #dataanalytics #firebolt #dataengineering

GOOD MORNING #babyreindeer #dataengineer #softwareengineer #developer

SUMMARY TABLES – YOU’RE WORKING HARDER THAN YOU NEED TO When dealing with large data volumes, summary tables are often used to accelerate query time performance by pre-aggregating. But instead of spending weeks moving from concept to production, you can try a different approach. Slow query? Give us 5 minutes, it's just admin index work: CREATE AGGREGATING INDEX ix_robs_happy_index on sometable (keyfield1, keyfield2, sum(someotherfield), avg(someotherfield)); And you’re done. Same result and the end user doesn't have to change the target of their query. All they know is that things are now 100x faster. And your compute bill drops. Speed isn't all about compute costs, it's also about manpower costs. If you can achieve the same result with a month less staff time, you're on to something. Robert Harmon #summarytables #indexing #database #datawarehouse #analytics

You're not providing anything if you can't create action. I don't care how many pipelines you created, how many dashboards you authored, how many bugs you fixed this week. As a data professional, if what you do doesn't cause action, it's worthless. Dashboards are passive. Sure, they're informative, that's great. They don't require ACTION. If I look at a dashboard measuring my performance or the performance of my team, I may or may not find a customer related problem. Even if I do, so long as my performance is roughly the same as my peers' performance, why would I change? I won't. The value of the whole system is almost 0. Compare this with a slack message that says something like: "Your team member Jeff is performing within the boundaries of his team, but we noticed his volume has increased from 10 customer interactions per day to 20 but his quality (customer satisfaction rate) has decreased from 90% to 85%, both by a half standard deviation from his historical norm. Please investigate, do what you can to correct this issue, then click on the following link to log your findings." Bang, just like that, we've induced action. We know Jeff's volume has increased, which is a good thing, but his customer sat rating has decreased, which is a bad thing. We need to talk to Jeff and figure out what's going on. The data warehouse already knows these things. It's fully capable of pointing out when things are not normal. It doesn't need AI or machine learning to do this. Why in the hell are we trying to fix it with dashboards, rather than inducing action? #datawarehouse #processmanagement

High data warehouse costs? It might be because your engines are always on. More and more platforms are offering a decoupled storage and compute architecture. Here’s why you should consider this feature in your next tech stack evaluation, specifically how Firebolt offers these capabilities in its data warehouse: 1. Online Scaling: Dynamically adjusts resources to accommodate peak hours, scaling data volumes, and concurrent usage. 2. On-Demand Capacity: Allows for quick resizing and flexible starting or stopping of engines. 3. Granular Incrementalities: Scales nodes from 1 to 128, ensuring you only use the resources necessary for the best price-performance ratio. Evaluating your current setup for these features could help you reduce costs and enhance efficiency. #datawarehouse #compute #storage