CassandraConf.ru program
Schedule and abstracts of reports are ready and available on the site CassandraConf.ru . So, what awaits us on December 9:
Foreign speakers:
')
Jonathan Ellis , DataStax Co-Founder and Apache Cassandra Project Manager
Apache Cassandra 2.0 and 2.1
This report will explain why it’s possible to make a lot of data, such as Adobe, CERN, eBay, Instagram, and Spotify. It will make it possible to understand the application of data modeling.
Cassandra plays in modern scale-out architectures. Cassandra 2.0 and the upcoming 2.1 release.
Axel Liljencrantz , backend developer of the Spotify music service
Cassandra performance. Underlying details of the compaction strategies
This is a concept of compaction. It is not a clear understanding of Cassandra performance.
Ruth Stryker , Technical Trainer in DataStax
CQL and the DataStax Java Driver
This is an introductory session to the CQL (Cassandra Query Language) and the Critical background. In this session, you will be able to learn how to use it. Then, you can learn how to install it.
Aleksey Yeschenko , Apache Cassandra committer, DataStax
Data Modeling with CQL3
The talk will deal with time series, collections, secondary indexes, CAS, and user types (2.1)
Nikolai Alimenkov , Technical Lead in ZoralLabs, trainer in XP Injection
The organization of high-loaded queues in Cassandra
Everyone knows that Cassandra is a NoSQL storage solution. But often, a queue system is used to process this data, which is implemented using other tools. Due to this, inconsistency in data and an additional level to maintain appear. Since for one of our products all data is stored in Cassandra, we developed our solution for the queue, which automatically acquired many useful qualities: scaling, high availability, flexibility. This decision will be presented in the report.
Track for experienced developers
Oleg Anastasyev , Odnoklassniki Lead Developer
Closer to Cassandra
The way to use Cassandra is a bit unusual - we don’t use thrift or netty-based native protocol to work with Cassandra using only the built-in query language, as with a normal DBMS. Instead, we merge the Cassandra node in the same JVM with the data warehouse business service. Clients use already business-oriented network requests to work with such a repository.
So we reduce the number of network calls and traffic. Also, it allows us to make small hacks of embedded data processing logic that allow you to build distributed data warehouses more efficiently and easily.
In the report we will talk about the practical application of this approach on the example of the three most interesting data warehouses in Odnoklassniki.
Olga Soboleva and Kirill Ivanov , SKB Kontur
Transaction processing on the example of a telecommunications company
The report on the example of the billing of the cellular operator will consider the service model, which is closely integrated with Cassandra. It will be described how the requirements have changed, how the system has evolved and how the architecture of Cassandra has changed. The report will be of interest to those who want to know how data storage problems are solved when using Cassandra.
Ivan Burmistrov and Fedor Fominykh , SKB Kontur
Strictly oriented sequence of time events
Cassandra is a highly scalable data warehouse in which all nodes are equal. This is a big plus for fault tolerance and ease of setup, but with such an architecture, the solution of some tasks becomes more complicated. This task includes the task of storing and retrieving objects in chronological order. This problem can be solved with the help of time series, which are well implemented in cassandra (you can read about them, for example, here ). However, the classic time series does not guarantee that at each moment in time all customers reading the objects of the time series will get the same result. This happens because the time of an object appearing in a time series is determined by the client. Therefore, an object with a longer time may be recorded earlier. But there are situations when the order of objects is important, and it is important to always return objects in the same order. The report will describe the modification of classical time series, devoid of this disadvantage. The report will be interesting to anyone who is interested in tasks that can be solved with the help of Cassandra.
Alexander Sabinin , Kaspersky Lab
Task Queuing for a Scanning Robot
As part of the activities of the content filtering laboratory, there is the task of maintaining up to date information about the domains visited by users of Kaspersky Lab products. The database of domains has several tens of millions of entries and is constantly updated. Information about each specific site is determined by specialized technologies. Obviously, the content of web resources can dynamically change, and this requires regular rescanning. For this reason, it became necessary to work out a specific set of heuristics for calculating the resource rescanning interval depending on the most significant factors. In addition, there is a need to quickly add tasks and manage the order of their execution. These factors do not allow to manage the fixed order of rescanning resources and require a more flexible approach.
When analyzing existing solutions, Cassandra was chosen as a reliable data storage with high bandwidth and providing flexibility for further data modifications. About exactly how Cassandra is used in Kaspersky Lab and will be discussed in the report.
Track for novice developers
Evgeny Tikhonov , SKB Kontur
Introduction to Cassandra
Cassandra is a relatively young data warehouse. In my report, I want to immerse the listeners in her device a little, to show how powerful and at the same time simple this tool is. From the report it will be clear to you how the data is distributed in a cluster, how the query coordinator works and what processes take place in one node while processing the request.
The report will be useful to beginners who have just started working with Cassandra, as well as those who want to briefly find out how the Cassandra cluster works in general terms.
Alexander Soloviev , ex GridDynamics contractor
Cassandra in e-commerce
The report will describe the process of transitioning a single project to data storage using Cassandra. To begin with, we will focus on the initial conditions in which the decision to transition was made by the high-level architecture, the reasons for migrating from the In-Memory Data Grid to Cassandra, the system performance requirements that we wanted to achieve. Then the steps that helped improve performance (TPS and / or latency) will be described. I will also tell about what did not lead to a significant improvement, and explain why some of our hypotheses did not work.
Maxim Sychev and Alexander Kokovin, SKB Kontur
How we moved to Cassandra
In the report we will tell why we decided to replace our storage system and why we chose Cassandra. We describe the difficulties encountered when replacing the storage and how to solve them. We will also dwell separately on the scenarios for the use of Cassandra, which we are currently practicing (task queue and data indexing on top of Cassandra).
Evgeny Kurpilyansky , SKB Kontur
Indexing over Cassandra
Cassandra is a highly scalable, reliable data warehouse. However, it does not allow to effectively perform complex search queries. The report will tell you how to index data from Cassandra in SQL-storage. This approach will allow you to perform arbitrary SQL queries on your data and at the same time preserve important properties of your system (such as scalability, fault tolerance, and others)
In the final of the conference, we plan 30 minutes for the favorite Lightning Talks format - short five-minute speeches without answers to questions, where everyone can share their case studies using Cassandra, technical solutions and urgent issues. If you already understand what you have to say, write to elena@it-people.ru.
After the official part - afterparty in the cafe InSalad (Gorky 63). It will be possible to talk with brothers in interests, ask questions to the speakers in an informal setting. It will be fun!
So again.
9th December. from 10.00 to 19.30.
Ekaterinburg, Novotel Engels Hotel Conference Hall 7.
Schedule and registration - on CassandraConf.ru
See you :)
ps Video recording of the reports will be publicly available a couple of weeks after the event.
Foreign speakers:
')
Jonathan Ellis , DataStax Co-Founder and Apache Cassandra Project Manager
Apache Cassandra 2.0 and 2.1
This report will explain why it’s possible to make a lot of data, such as Adobe, CERN, eBay, Instagram, and Spotify. It will make it possible to understand the application of data modeling.
Cassandra plays in modern scale-out architectures. Cassandra 2.0 and the upcoming 2.1 release.
Axel Liljencrantz , backend developer of the Spotify music service
Cassandra performance. Underlying details of the compaction strategies
This is a concept of compaction. It is not a clear understanding of Cassandra performance.
Ruth Stryker , Technical Trainer in DataStax
CQL and the DataStax Java Driver
This is an introductory session to the CQL (Cassandra Query Language) and the Critical background. In this session, you will be able to learn how to use it. Then, you can learn how to install it.
Aleksey Yeschenko , Apache Cassandra committer, DataStax
Data Modeling with CQL3
The talk will deal with time series, collections, secondary indexes, CAS, and user types (2.1)
Nikolai Alimenkov , Technical Lead in ZoralLabs, trainer in XP Injection
The organization of high-loaded queues in Cassandra
Everyone knows that Cassandra is a NoSQL storage solution. But often, a queue system is used to process this data, which is implemented using other tools. Due to this, inconsistency in data and an additional level to maintain appear. Since for one of our products all data is stored in Cassandra, we developed our solution for the queue, which automatically acquired many useful qualities: scaling, high availability, flexibility. This decision will be presented in the report.
Track for experienced developers
Oleg Anastasyev , Odnoklassniki Lead Developer
Closer to Cassandra
The way to use Cassandra is a bit unusual - we don’t use thrift or netty-based native protocol to work with Cassandra using only the built-in query language, as with a normal DBMS. Instead, we merge the Cassandra node in the same JVM with the data warehouse business service. Clients use already business-oriented network requests to work with such a repository.
So we reduce the number of network calls and traffic. Also, it allows us to make small hacks of embedded data processing logic that allow you to build distributed data warehouses more efficiently and easily.
In the report we will talk about the practical application of this approach on the example of the three most interesting data warehouses in Odnoklassniki.
Olga Soboleva and Kirill Ivanov , SKB Kontur
Transaction processing on the example of a telecommunications company
The report on the example of the billing of the cellular operator will consider the service model, which is closely integrated with Cassandra. It will be described how the requirements have changed, how the system has evolved and how the architecture of Cassandra has changed. The report will be of interest to those who want to know how data storage problems are solved when using Cassandra.
Ivan Burmistrov and Fedor Fominykh , SKB Kontur
Strictly oriented sequence of time events
Cassandra is a highly scalable data warehouse in which all nodes are equal. This is a big plus for fault tolerance and ease of setup, but with such an architecture, the solution of some tasks becomes more complicated. This task includes the task of storing and retrieving objects in chronological order. This problem can be solved with the help of time series, which are well implemented in cassandra (you can read about them, for example, here ). However, the classic time series does not guarantee that at each moment in time all customers reading the objects of the time series will get the same result. This happens because the time of an object appearing in a time series is determined by the client. Therefore, an object with a longer time may be recorded earlier. But there are situations when the order of objects is important, and it is important to always return objects in the same order. The report will describe the modification of classical time series, devoid of this disadvantage. The report will be interesting to anyone who is interested in tasks that can be solved with the help of Cassandra.
Alexander Sabinin , Kaspersky Lab
Task Queuing for a Scanning Robot
As part of the activities of the content filtering laboratory, there is the task of maintaining up to date information about the domains visited by users of Kaspersky Lab products. The database of domains has several tens of millions of entries and is constantly updated. Information about each specific site is determined by specialized technologies. Obviously, the content of web resources can dynamically change, and this requires regular rescanning. For this reason, it became necessary to work out a specific set of heuristics for calculating the resource rescanning interval depending on the most significant factors. In addition, there is a need to quickly add tasks and manage the order of their execution. These factors do not allow to manage the fixed order of rescanning resources and require a more flexible approach.
When analyzing existing solutions, Cassandra was chosen as a reliable data storage with high bandwidth and providing flexibility for further data modifications. About exactly how Cassandra is used in Kaspersky Lab and will be discussed in the report.
Track for novice developers
Evgeny Tikhonov , SKB Kontur
Introduction to Cassandra
Cassandra is a relatively young data warehouse. In my report, I want to immerse the listeners in her device a little, to show how powerful and at the same time simple this tool is. From the report it will be clear to you how the data is distributed in a cluster, how the query coordinator works and what processes take place in one node while processing the request.
The report will be useful to beginners who have just started working with Cassandra, as well as those who want to briefly find out how the Cassandra cluster works in general terms.
Alexander Soloviev , ex GridDynamics contractor
Cassandra in e-commerce
The report will describe the process of transitioning a single project to data storage using Cassandra. To begin with, we will focus on the initial conditions in which the decision to transition was made by the high-level architecture, the reasons for migrating from the In-Memory Data Grid to Cassandra, the system performance requirements that we wanted to achieve. Then the steps that helped improve performance (TPS and / or latency) will be described. I will also tell about what did not lead to a significant improvement, and explain why some of our hypotheses did not work.
Maxim Sychev and Alexander Kokovin, SKB Kontur
How we moved to Cassandra
In the report we will tell why we decided to replace our storage system and why we chose Cassandra. We describe the difficulties encountered when replacing the storage and how to solve them. We will also dwell separately on the scenarios for the use of Cassandra, which we are currently practicing (task queue and data indexing on top of Cassandra).
Evgeny Kurpilyansky , SKB Kontur
Indexing over Cassandra
Cassandra is a highly scalable, reliable data warehouse. However, it does not allow to effectively perform complex search queries. The report will tell you how to index data from Cassandra in SQL-storage. This approach will allow you to perform arbitrary SQL queries on your data and at the same time preserve important properties of your system (such as scalability, fault tolerance, and others)
In the final of the conference, we plan 30 minutes for the favorite Lightning Talks format - short five-minute speeches without answers to questions, where everyone can share their case studies using Cassandra, technical solutions and urgent issues. If you already understand what you have to say, write to elena@it-people.ru.
After the official part - afterparty in the cafe InSalad (Gorky 63). It will be possible to talk with brothers in interests, ask questions to the speakers in an informal setting. It will be fun!
So again.
9th December. from 10.00 to 19.30.
Ekaterinburg, Novotel Engels Hotel Conference Hall 7.
Schedule and registration - on CassandraConf.ru
See you :)
ps Video recording of the reports will be publicly available a couple of weeks after the event.
Source: https://habr.com/ru/post/204648/
All Articles