Nick Gerner interviewed at Velocity 2010
summary
Open-source and low-cost monitoring tools exist but many fail at handling dynamic nodes as cloud instances appear and disappear; provider offerings vary widely — instance types split CPU, I/O and pricing, making performance characteristics nontrivial to predict. Cloud is cost-effective for small teams but can become more expensive than owning hardware at enterprise scale. A cited S3 reliability claim argues data-loss rates are vanishingly small. Google treats page load time as a ranking signal and provides concrete Webmaster Tools measurements and remediation suggestions.
AI-written summary · about
related talks
50mThe Great Open Cloud Shootout
17mCloud Compute Cost Showdown
11mFlip Kromer interviewed at Velocity 2011
42mData Driven Scalability and Cost Analysis for Evolutionary Architecture In the Cloud
36m3 Pitfalls Everyone Should Avoid with Cloud Data
38mWhat's Next after Cloud Meets Big Data
44mOpen Source Analytics, Public Clouds, and VC Mega-Investment
59mBuilding the Best Cloud Data Architecture: Avoiding Costly Pitfalls