Vertica on HighLoad ++
Yesterday was my performance on HighLoad ++. Abstracts and slides on the website of the organizers . The conference is organized, by the way, perfectly. But time for a full-fledged performance was not enough - 45 minutes with questions. The test run took me 60 minutes, after some reorganization and no questions on HL, I managed to complete 42. Some important architectural points had to be pronounced quickly and without examples, from which, of course, clarity suffered. I tried to build the presentation in such a way as to show how we came to Vertical and the current architecture, while at the same time focusing on important architectural principles of working with big data in general. Not sure that the goal was fully achieved. Little, little time. But I am always open to questions. Vertika, however, caused a well-deserved interest, the questions were in the case.
And today there was a speech by Chris Bonn from etsy.com, and, surprisingly, he also talked about Vertica. Two reports on HL is already a trend. Chris tried to tell in another way what the RDBMS was, spent almost half an hour on this and did not have time either. However, for my taste, they too chewed some things, but I am subjective. Compared with my report yesterday, he made two useful additions to the audience. Firstly, he rightly pointed out that the one-port Vertika up to 1TB is free. Although it is not enough, it is much better than any other single-ended RDBMS for analytics. Secondly, he tried to make a demo. I did not guess, but I did not have access to the computer during the performance. However, his demo did not come out very well, and the example of the “quick question” was not very indicative, since it was not clear how many rows there are in the table. It was necessary to take a table of several billion lines, and make a query with the aggregation of all data for a couple of seconds on the laptop. An example is not difficult to pick up, but it would be very impressive.
It is possible that due to these reports, interest in Vertic will grow in our country. If someone saw these performances in HL ++, your impressions are interesting, that it was clear that not.
And today there was a speech by Chris Bonn from etsy.com, and, surprisingly, he also talked about Vertica. Two reports on HL is already a trend. Chris tried to tell in another way what the RDBMS was, spent almost half an hour on this and did not have time either. However, for my taste, they too chewed some things, but I am subjective. Compared with my report yesterday, he made two useful additions to the audience. Firstly, he rightly pointed out that the one-port Vertika up to 1TB is free. Although it is not enough, it is much better than any other single-ended RDBMS for analytics. Secondly, he tried to make a demo. I did not guess, but I did not have access to the computer during the performance. However, his demo did not come out very well, and the example of the “quick question” was not very indicative, since it was not clear how many rows there are in the table. It was necessary to take a table of several billion lines, and make a query with the aggregation of all data for a couple of seconds on the laptop. An example is not difficult to pick up, but it would be very impressive.
It is possible that due to these reports, interest in Vertic will grow in our country. If someone saw these performances in HL ++, your impressions are interesting, that it was clear that not.
')
Source: https://habr.com/ru/post/155773/
All Articles