Hi, Here are my comments.
1. It's MapReduce, however you can start several nodes per physical box to better scale the execution across available CPUs. 2. Data in Ignite is stored in key-value pairs which is completely different approach. But if to choose between the two, it's much closer to row based storage. 3. This thread has a good discussion about this: http://stackoverflow.com/questions/33936603/apache-ignite-vs-apache-storm-in-depth 4. Yes, you can span the cluster across several data centers, but this will most likely introduce increased latencies and performance drop. However, there are third party offers for async replication, like the one provided by GridGain in their paid solution: https://gridgain.readme.io/docs/data-center-replication -Val -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/where-to-get-the-document-about-Ignite-architecture-or-technical-advantage-tp8282p8309.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.