SlideShare a Scribd company logo
InfiniFlux vs RDBMS
www.infiniflux.com
Overview
2
• The concept of InfiniFlux: the ultra high-speed database that stores and processes time series data.
• InfiniFlux has very different characteristics compared to the conventional DBMS such as Oracle and DB2 in
order to provide high-speed processing.
• Importance of understanding: the characteristics and architecture for time series log data.
• Technical characteristics: analyzing data using SQL and storing hundreds of thousands of records in a
second in real-time.
• Describe the differences between InfiniFlux and the conventional technology, and provide detailed
explanation about every item.
Document Overview
Comparison Chart
3
Characteristics InfiniFlux DB RDBMS
Transaction
Provide implicit transaction based on
Snapshot
Provide explicit transaction based on log file
INSERT speed 300,000 ~ 3,000,000 in a second less than 10,000 in a second
SELECT performance
Optimized for query over search & statistical
analysis
Optimized for OLTP
Characteristics of Data Time series log data Non-log transaction data
Updatable data Append Only (Write Once, Read Many) Updatable
DELETE operation Delete the oldest data Delete random data
Real-time index Real-time bitmap index Not real-time B+Tree
Data compression
High performance data compression in real-
time
Not support data compression
Real-time search for text Supported (real-time inverted index)
Not supported
(Even if supported, not in real-time)
Support time series data Support data partitioning by sharding
Support data partitioning based on general
timestamp column
Gap between data and index Occurrence of momentary gap Gap not occurred
Support Transaction
4
RDBMS
• Provide explicit transaction for all the data operation
• Transaction: set of operations that satisfy ACID properties
• Atomicity
• Consistency
• Isolation
• Durability
• Savepoint, Commit, and Rollback
InfiniFlux
• Not provide explicit transaction
• No transaction for data operation (input)
• Provide implicit transaction over internal meta data
• Table structure, index structure, and data file structure
Reasons for not
supporting transaction
• No need to store data based on transaction since it is time series data
• Fast storage and processing are much more valuable than transaction
• No need to pay costs and conduct complex operations for transactions
Input Performance
5
RDBMS
• Difficult to input data more than 5,000 per second
• High costs of logging for transaction. As a result, increase I/O costs.
• High costs of index update. Thus, B+Tree is suitable for data search.
• To maintain the consistency, all the operations of index update for a record is conducted in
consecutive order.
• Degrading system performance as the volume of index is greater than data
InfiniFlux
• Able to input hundreds of thousands data per second
• Efficient to create index through real-time bitmap index
• Costs for logging not required
• Parallel index can be created using multi-threads
• Reduce the amount of I/O data by real-time compression. As a result, improve overall system
performance.
• Able to improve performance greatly by creating tablespace based on multiple disks
Query Performance
6
RDBMS
• Row-oriented database has advantage over online transaction processing (OLTP).
• High query performance on high cardinality
• Reason: small search range brings high performance like B+Tree
• Mainly operating based on B+Tree, and select the most efficient index and use it.
• Efficient to search in a small range even with large number of data
• Slow query performance for statistical analysis
InfiniFlux
• Column-oriented database has advantage over online analytical processing (OLAP).
• High query performance on low cardinality
• Most of time series log data are low cardinality since it has high level of duplication.
• Most of them are bitmap indices, and it is efficient since more than two indices can be used at
the same time.
• Fast statistical query against massive data (hundreds of millions of data)
• Relatively slow on search for a certain record of OLTP from the whole DB
• In this case, global index needs to be created.
Characteristics of Data
7
RDBMS
• Optimized for storing data through transactions
• Financial information and individual identification information for banking transactions
• Important data that should be safely stored in conventional database such as Oracle.
• RDMBS is not related with the flow of time and can be updated or deleted.
InfiniFlux
• Target data are time series log data.
• Hundreds of thousands of data were created in a second.
• Update operation is not required and shows high level of data duplication.
• Target data are log files or similar data that were previously stored as text files.
• Constantly describe the status of a certain target over time.
Updatable Data
8
RDBMS
• Updatable model
• Data can be updated or modified anytime.
• Able to delete an arbitrary record at a random moment.
• Create a database that can access and modify all the data with ease.
InfiniFlux
• Write Once Read Many (WORM) model
• Since it is time series log data, data cannot be modified once it is stored.
• Read-oriented, and not support UPDATE operation at all
• Deleting data?
• Cannot delete a record at a certain time
• Able to delete the oldest record from the database in sequential order
* InfiniFlux provides delete feature in order to maintain a certain level of disk usage of
embedded device.
Index Technology
9
RDBMS
• In general, it uses B+Tree
• It is global index and all the record information are stored in the index.
• Expensive costs of logging and recovery operations for supporting transactions
* Due to the reasons mentioned above, it process only thousands data per second.
• Disk usage is increasing rapidly compared to the number of indices because values of raw data are
stored in the index.
• Not provide compression feature due to performance issue
InfiniFlux
• Support real-time bitmap index
• Local index structure of partition unit
• Composed with various indices over all the records
• Able to create index quickly since transaction support and logging operations are not required.
• Able to create millions of indices per second
• Costs of creating index will increase little by little even though the amount of data increased.
• Minimize the disk usage by supporting compression algorithms
• OLTP query is relatively slow when the amount of data increases indefinitely.
• Support semi-global index (4th quarter 2015)
• Able to provide high performance on OLTP query
• Able to provide high performance on compression and statistical query
Compression Technology
10
RDBMS
• No data compression issues.
* The reason for this is that, in general, the number of target records are maintained at a certain number.
In this environment, disk usage is not that big issue.
• The basic idea is that increase search performance at the expense of disk space.
• In many cases, clients are not interested in data compression itself.
• Even if compression is available, challenges below will be presented;
• Increase the amount of data usage due to the structure in data duplication of B+Tree
• Difficult to use the duplication property of row-oriented database
InfiniFlux
• Amount of disk usage is very important since it is assumed that the environment where large amount of
data is created.
• Conduct real-time compression twice and thus able to use disk space efficiently
• Logical compression
• Compress column-based duplicated data in a dictionary structure
• Duplicated bit strings over bitmap index at high compression logically
• Physical compression
• Real-time physical compression of partition pages when storing disks
• Disk usages show slow growth pattern not linear growth as the number of index increases.
Full Text Search
11
RDBMS
• In general, database do not support text search
• Provide alternative methods such as LIKE statement for the search.
• Partial search over the column were conducted through LIKE ‘%pattern%’.
In this case, full scan over all the records is operated, and resulted in slow performance.
• Impossible to use RDBMS for full text search
InfiniFlux
• Provide keyword index, and full text search is available for a certain VARCHAR column.
• “SEARCH” can be used for searching instead of “LIKE”.
• SELECT id FROM user_table WHERE address SEARCH ‘Texas’;
• Able to search text in UTF-8 format.
• Different processing methods are applied to English characters and 2 byte character
(Chinese/Japanese/Korean).
• English
• Separate words with special characters (e.g. “boy” and “boys” are different)
• CJK (Chinese, Japanese, and Korean)
• Using 2-gram method to search
• The word ‘강남구’ will be converted into ‘강남’ & ‘남구’ and operate
Concurrency Level
12
RDBMS
• Generally provide “Record Level Locking”
• Two options based on how to read records while updating
• Consistency Read
• Allow to read previous value of the record that is being updated.
• No confliction of lock  No waiting time
• Non Consistency Read (Record locking conflict occurred)
• Allow to wait and decide to commit/rollback on the record that is being updated.
• Conflicts arise while locking  Possible to wait indefinitely depends on the previous
transaction.
InfiniFlux
• Provide Lockless structure
• No locking conflict due to no update
• No lock mechanism against the records, thus, maximize the performance over data search and
analysis.
Time Series Data Analysis
13
Sliding
Memory
Window
Memory
Window
Memory
Window
File -1 File -2 File -3 File -4 File -5 File -6 File -7
Data
Insert
Current time Old time
RDBMS
• Time series data exist in the form of time data type (date, time, timestamp, and interval).
• It doesn’t recognize the data as time series data, rather treat it with other general data (number,
and varchar etc.).
• Separate devices for analyzing data based on the time base is not provided due to the reasons
above.
• Slow performance when analyzing data with time index.
InfiniFlux
• When data were entered, it creates physical partitions in sequential order.
• Able to access records in a certain time range directly due to the sequential order.
• When data were entered, the timestamp for the record will be stored in nano unit automatically
(_arrival_time as a hidden column).
• Able to operate data of the time base freely based on the hidden column.
• Select data from table where DURATION 10 minute (output data within 10 minutes)
Backup and Restore
14
Backup Restore
RDBMS
Backup data in a certain table or database area
externally
• Online backup is the basic strategy to store data
• Separately provide incremental backup in order to
reduce the time and costs of backup
The process of returning the data to the original database
• Impossible to use backup files and must go through
restore process
• Backup big files take a long time to restore
• Able to backup data based on a certain time, but
complicate to change the starting point.
InfiniFlux
• Backup the whole database based on a certain time
• Not able to backup data based on a certain table or
record
• Backup in a file or directory format multiple files
• Restore the whole backup files
• Overwrite the existing database
Backup and Mount
15
What is mount?
• Rather than storing backup database file, it loads and searches the contents of database after
loading only meta data.
• Similar to the concept of the disk mount of UNIX
• Able to access backup data directly in a certain time in a fast and efficient way.
RDBMS • Not supported
InfiniFlux
• MOUNT
• Mount backup data on 31st December 2014
e.g.) MOUNT DATABASE ‘/home/data/2014-12-31’
• UNMOUNT
• Unmount the mounted data
e.g.) UNMOUNT DATABASE ‘/home/data/2014-12-31’
The World's Fastest
Time Series DBMS
for IoT and Big Data
www.infiniflux.com
info@infiniflux.com
InfiniFlux

More Related Content

What's hot

MongoDB Capacity Planning
MongoDB Capacity PlanningMongoDB Capacity Planning
MongoDB Capacity Planning
Norberto Leite
 
Hardware Provisioning for MongoDB
Hardware Provisioning for MongoDBHardware Provisioning for MongoDB
Hardware Provisioning for MongoDBMongoDB
 
Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...
Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...
Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...
Alluxio, Inc.
 
Hybrid collaborative tiered storage with alluxio
Hybrid collaborative tiered storage with alluxioHybrid collaborative tiered storage with alluxio
Hybrid collaborative tiered storage with alluxio
Thai Bui
 
Presto updates to 0.178
Presto updates to 0.178Presto updates to 0.178
Presto updates to 0.178
Kai Sasaki
 
MongoDB Replication fundamentals - Desert Code Camp - October 2014
MongoDB Replication fundamentals - Desert Code Camp - October 2014MongoDB Replication fundamentals - Desert Code Camp - October 2014
MongoDB Replication fundamentals - Desert Code Camp - October 2014
Avinash Ramineni
 
NoSQL for SQL Users
NoSQL for SQL UsersNoSQL for SQL Users
NoSQL for SQL Users
IBM Cloud Data Services
 
When to Use MongoDB...and When You Should Not...
When to Use MongoDB...and When You Should Not...When to Use MongoDB...and When You Should Not...
When to Use MongoDB...and When You Should Not...
MongoDB
 
MongoDB vs Mysql. A devops point of view
MongoDB vs Mysql. A devops point of viewMongoDB vs Mysql. A devops point of view
MongoDB vs Mysql. A devops point of view
Pierre Baillet
 
Aerospike Hybrid Memory Architecture
Aerospike Hybrid Memory ArchitectureAerospike Hybrid Memory Architecture
Aerospike Hybrid Memory Architecture
Aerospike, Inc.
 
Why You Definitely Don’t Want to Build Your Own Time Series Database
Why You Definitely Don’t Want to Build Your Own Time Series DatabaseWhy You Definitely Don’t Want to Build Your Own Time Series Database
Why You Definitely Don’t Want to Build Your Own Time Series Database
InfluxData
 
Сергей Сверчков и Виталий Руденя. Choosing a NoSQL database
Сергей Сверчков и Виталий Руденя. Choosing a NoSQL databaseСергей Сверчков и Виталий Руденя. Choosing a NoSQL database
Сергей Сверчков и Виталий Руденя. Choosing a NoSQL database
Volha Banadyseva
 
Hadoop 2 cluster architecture
Hadoop 2 cluster architectureHadoop 2 cluster architecture
Hadoop 2 cluster architecture
Sandeep Patil
 
MongoDB Administration ~ Kevin Hanson
MongoDB Administration ~ Kevin HansonMongoDB Administration ~ Kevin Hanson
MongoDB Administration ~ Kevin Hanson
hungarianhc
 
In-Memory Computing: How, Why? and common Patterns
In-Memory Computing: How, Why? and common PatternsIn-Memory Computing: How, Why? and common Patterns
In-Memory Computing: How, Why? and common Patterns
Srinath Perera
 
Capacity Planning For Your Growing MongoDB Cluster
Capacity Planning For Your Growing MongoDB ClusterCapacity Planning For Your Growing MongoDB Cluster
Capacity Planning For Your Growing MongoDB Cluster
MongoDB
 
Ops Jumpstart: MongoDB Administration 101
Ops Jumpstart: MongoDB Administration 101Ops Jumpstart: MongoDB Administration 101
Ops Jumpstart: MongoDB Administration 101
MongoDB
 
How to ensure Presto scalability 
in multi use case
How to ensure Presto scalability 
in multi use case How to ensure Presto scalability 
in multi use case
How to ensure Presto scalability 
in multi use case
Kai Sasaki
 
Presto at Twitter
Presto at TwitterPresto at Twitter
Presto at Twitter
Bill Graham
 
RedisConf17 - Home Depot - Turbo charging existing applications with Redis
RedisConf17 - Home Depot - Turbo charging existing applications with RedisRedisConf17 - Home Depot - Turbo charging existing applications with Redis
RedisConf17 - Home Depot - Turbo charging existing applications with Redis
Redis Labs
 

What's hot (20)

MongoDB Capacity Planning
MongoDB Capacity PlanningMongoDB Capacity Planning
MongoDB Capacity Planning
 
Hardware Provisioning for MongoDB
Hardware Provisioning for MongoDBHardware Provisioning for MongoDB
Hardware Provisioning for MongoDB
 
Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...
Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...
Optimizing Latency-Sensitive Queries for Presto at Facebook: A Collaboration ...
 
Hybrid collaborative tiered storage with alluxio
Hybrid collaborative tiered storage with alluxioHybrid collaborative tiered storage with alluxio
Hybrid collaborative tiered storage with alluxio
 
Presto updates to 0.178
Presto updates to 0.178Presto updates to 0.178
Presto updates to 0.178
 
MongoDB Replication fundamentals - Desert Code Camp - October 2014
MongoDB Replication fundamentals - Desert Code Camp - October 2014MongoDB Replication fundamentals - Desert Code Camp - October 2014
MongoDB Replication fundamentals - Desert Code Camp - October 2014
 
NoSQL for SQL Users
NoSQL for SQL UsersNoSQL for SQL Users
NoSQL for SQL Users
 
When to Use MongoDB...and When You Should Not...
When to Use MongoDB...and When You Should Not...When to Use MongoDB...and When You Should Not...
When to Use MongoDB...and When You Should Not...
 
MongoDB vs Mysql. A devops point of view
MongoDB vs Mysql. A devops point of viewMongoDB vs Mysql. A devops point of view
MongoDB vs Mysql. A devops point of view
 
Aerospike Hybrid Memory Architecture
Aerospike Hybrid Memory ArchitectureAerospike Hybrid Memory Architecture
Aerospike Hybrid Memory Architecture
 
Why You Definitely Don’t Want to Build Your Own Time Series Database
Why You Definitely Don’t Want to Build Your Own Time Series DatabaseWhy You Definitely Don’t Want to Build Your Own Time Series Database
Why You Definitely Don’t Want to Build Your Own Time Series Database
 
Сергей Сверчков и Виталий Руденя. Choosing a NoSQL database
Сергей Сверчков и Виталий Руденя. Choosing a NoSQL databaseСергей Сверчков и Виталий Руденя. Choosing a NoSQL database
Сергей Сверчков и Виталий Руденя. Choosing a NoSQL database
 
Hadoop 2 cluster architecture
Hadoop 2 cluster architectureHadoop 2 cluster architecture
Hadoop 2 cluster architecture
 
MongoDB Administration ~ Kevin Hanson
MongoDB Administration ~ Kevin HansonMongoDB Administration ~ Kevin Hanson
MongoDB Administration ~ Kevin Hanson
 
In-Memory Computing: How, Why? and common Patterns
In-Memory Computing: How, Why? and common PatternsIn-Memory Computing: How, Why? and common Patterns
In-Memory Computing: How, Why? and common Patterns
 
Capacity Planning For Your Growing MongoDB Cluster
Capacity Planning For Your Growing MongoDB ClusterCapacity Planning For Your Growing MongoDB Cluster
Capacity Planning For Your Growing MongoDB Cluster
 
Ops Jumpstart: MongoDB Administration 101
Ops Jumpstart: MongoDB Administration 101Ops Jumpstart: MongoDB Administration 101
Ops Jumpstart: MongoDB Administration 101
 
How to ensure Presto scalability 
in multi use case
How to ensure Presto scalability 
in multi use case How to ensure Presto scalability 
in multi use case
How to ensure Presto scalability 
in multi use case
 
Presto at Twitter
Presto at TwitterPresto at Twitter
Presto at Twitter
 
RedisConf17 - Home Depot - Turbo charging existing applications with Redis
RedisConf17 - Home Depot - Turbo charging existing applications with RedisRedisConf17 - Home Depot - Turbo charging existing applications with Redis
RedisConf17 - Home Depot - Turbo charging existing applications with Redis
 

Viewers also liked

Relational database management systems
Relational database management systemsRelational database management systems
Relational database management systems
Datasoft Consulting
 
UNM-Taos Library
UNM-Taos LibraryUNM-Taos Library
UNM-Taos Library3nrico
 
Single Proprietorship JD
Single Proprietorship JDSingle Proprietorship JD
Single Proprietorship JDDean Apao
 
2015 znt inverter catalouge
2015 znt inverter catalouge2015 znt inverter catalouge
2015 znt inverter catalougeYe Arlene
 
SQL Server 2012 aamiaisseminaari
SQL Server 2012 aamiaisseminaariSQL Server 2012 aamiaisseminaari
SQL Server 2012 aamiaisseminaari
Salcom Group
 
La narración
La narraciónLa narración
La narración
Pedro Blanco Salceda
 
70-646 Windows Server 2008 Server Administrator
 70-646 Windows Server 2008 Server Administrator 70-646 Windows Server 2008 Server Administrator
70-646 Windows Server 2008 Server Administrator
maefrova
 
Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)
Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)
Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)
Nurul Aqila Ahamad Kamal
 
Substationoverview
Substationoverview Substationoverview
Substationoverview sailusree
 
InfiniFlux duration
InfiniFlux durationInfiniFlux duration
InfiniFlux duration
InfiniFlux
 
InfiniFlux Minmax Cache
InfiniFlux Minmax CacheInfiniFlux Minmax Cache
InfiniFlux Minmax Cache
InfiniFlux
 
TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...
TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...
TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...
tdc-globalcode
 
InfiniFlux collector
InfiniFlux collectorInfiniFlux collector
InfiniFlux collector
InfiniFlux
 
شهادة خبرة مكتب الهيكل العمرانى
شهادة خبرة مكتب الهيكل العمرانىشهادة خبرة مكتب الهيكل العمرانى
شهادة خبرة مكتب الهيكل العمرانى
Wafik Taha
 
Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2
Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2
Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2
Andrew Sungjin Kim
 
InfiniFlux IP Address Type
InfiniFlux IP Address TypeInfiniFlux IP Address Type
InfiniFlux IP Address Type
InfiniFlux
 
InfiniFlux performance
InfiniFlux performanceInfiniFlux performance
InfiniFlux performance
InfiniFlux
 
InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2
InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2
InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2
Andrew Sungjin Kim
 

Viewers also liked (20)

Relational database management systems
Relational database management systemsRelational database management systems
Relational database management systems
 
UNM-Taos Library
UNM-Taos LibraryUNM-Taos Library
UNM-Taos Library
 
Vusi Sibisi CV 2014
Vusi Sibisi CV 2014Vusi Sibisi CV 2014
Vusi Sibisi CV 2014
 
Single Proprietorship JD
Single Proprietorship JDSingle Proprietorship JD
Single Proprietorship JD
 
2015 znt inverter catalouge
2015 znt inverter catalouge2015 znt inverter catalouge
2015 znt inverter catalouge
 
ProTrac For Wholesalers
ProTrac For WholesalersProTrac For Wholesalers
ProTrac For Wholesalers
 
SQL Server 2012 aamiaisseminaari
SQL Server 2012 aamiaisseminaariSQL Server 2012 aamiaisseminaari
SQL Server 2012 aamiaisseminaari
 
La narración
La narraciónLa narración
La narración
 
70-646 Windows Server 2008 Server Administrator
 70-646 Windows Server 2008 Server Administrator 70-646 Windows Server 2008 Server Administrator
70-646 Windows Server 2008 Server Administrator
 
Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)
Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)
Heritage Studies. Proposal Slides (Japan & Australia) (Sem 1, 2015/2016)
 
Substationoverview
Substationoverview Substationoverview
Substationoverview
 
InfiniFlux duration
InfiniFlux durationInfiniFlux duration
InfiniFlux duration
 
InfiniFlux Minmax Cache
InfiniFlux Minmax CacheInfiniFlux Minmax Cache
InfiniFlux Minmax Cache
 
TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...
TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...
TDC2016POA | Trilha Bigdata - Armazenando séries temporais em bases de dados ...
 
InfiniFlux collector
InfiniFlux collectorInfiniFlux collector
InfiniFlux collector
 
شهادة خبرة مكتب الهيكل العمرانى
شهادة خبرة مكتب الهيكل العمرانىشهادة خبرة مكتب الهيكل العمرانى
شهادة خبرة مكتب الهيكل العمرانى
 
Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2
Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2
Infiniflux vs influxdb 비교 테스트 결과 2016 12월-v2
 
InfiniFlux IP Address Type
InfiniFlux IP Address TypeInfiniFlux IP Address Type
InfiniFlux IP Address Type
 
InfiniFlux performance
InfiniFlux performanceInfiniFlux performance
InfiniFlux performance
 
InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2
InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2
InfiniFlux vs influxdb 비교 테스트 결과 2016 12월-v2
 

Similar to InfiniFlux vs_RDBMS

Best storage engine for MySQL
Best storage engine for MySQLBest storage engine for MySQL
Best storage engine for MySQL
tomflemingh2
 
AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...
AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...
AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...
Amazon Web Services
 
Database Technologies
Database TechnologiesDatabase Technologies
Database Technologies
Michel de Goede
 
Introduction to Azure DocumentDB
Introduction to Azure DocumentDBIntroduction to Azure DocumentDB
Introduction to Azure DocumentDB
Radenko Zec
 
Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...
Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...
Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...Maaz Anjum
 
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostLeverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Zilliz
 
week1slides1704202828322.pdf
week1slides1704202828322.pdfweek1slides1704202828322.pdf
week1slides1704202828322.pdf
TusharAgarwal49094
 
Data Kinetics Products
Data Kinetics ProductsData Kinetics Products
Data Kinetics Products
sheena82
 
Redshift overview
Redshift overviewRedshift overview
Redshift overview
Amazon Web Services LATAM
 
Data Virtualization Reference Architectures: Correctly Architecting your Solu...
Data Virtualization Reference Architectures: Correctly Architecting your Solu...Data Virtualization Reference Architectures: Correctly Architecting your Solu...
Data Virtualization Reference Architectures: Correctly Architecting your Solu...
Denodo
 
Still All on One Server: Perforce at Scale
Still All on One Server: Perforce at Scale Still All on One Server: Perforce at Scale
Still All on One Server: Perforce at Scale
Perforce
 
Elastic storage in the cloud session 5224 final v2
Elastic storage in the cloud session 5224 final v2Elastic storage in the cloud session 5224 final v2
Elastic storage in the cloud session 5224 final v2
BradDesAulniers2
 
Gruter TECHDAY 2014 Realtime Processing in Telco
Gruter TECHDAY 2014 Realtime Processing in TelcoGruter TECHDAY 2014 Realtime Processing in Telco
Gruter TECHDAY 2014 Realtime Processing in Telco
Gruter
 
Comparative study of modern databases
Comparative study of modern databasesComparative study of modern databases
Comparative study of modern databases
Anirban Konar
 
AWS Cloud experience concepts tips and tricks
AWS Cloud experience concepts tips and tricksAWS Cloud experience concepts tips and tricks
AWS Cloud experience concepts tips and tricks
Dirk Harms-Merbitz
 
S016828 storage-tiering-nola-v1710b
S016828 storage-tiering-nola-v1710bS016828 storage-tiering-nola-v1710b
S016828 storage-tiering-nola-v1710b
Tony Pearson
 
Best Practices – Extreme Performance with Data Warehousing on Oracle Databa...
Best Practices –  Extreme Performance with Data Warehousing  on Oracle Databa...Best Practices –  Extreme Performance with Data Warehousing  on Oracle Databa...
Best Practices – Extreme Performance with Data Warehousing on Oracle Databa...
Edgar Alejandro Villegas
 
A12 vercelletto indexing_techniques
A12 vercelletto indexing_techniquesA12 vercelletto indexing_techniques
A12 vercelletto indexing_techniques
BeGooden-IT Consulting
 
Lecture 05 - The Data Warehouse and Technology
Lecture 05 - The Data Warehouse and TechnologyLecture 05 - The Data Warehouse and Technology
Lecture 05 - The Data Warehouse and Technology
phanleson
 
Slides: Start Small, Grow Big with a Unified Scale-Out Infrastructure
Slides: Start Small, Grow Big with a Unified Scale-Out InfrastructureSlides: Start Small, Grow Big with a Unified Scale-Out Infrastructure
Slides: Start Small, Grow Big with a Unified Scale-Out Infrastructure
NetApp
 

Similar to InfiniFlux vs_RDBMS (20)

Best storage engine for MySQL
Best storage engine for MySQLBest storage engine for MySQL
Best storage engine for MySQL
 
AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...
AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...
AWS re:Invent 2016| DAT318 | Migrating from RDBMS to NoSQL: How Sony Moved fr...
 
Database Technologies
Database TechnologiesDatabase Technologies
Database Technologies
 
Introduction to Azure DocumentDB
Introduction to Azure DocumentDBIntroduction to Azure DocumentDB
Introduction to Azure DocumentDB
 
Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...
Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...
Maaz Anjum - IOUG Collaborate 2013 - An Insight into Space Realization on ODA...
 
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostLeverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
 
week1slides1704202828322.pdf
week1slides1704202828322.pdfweek1slides1704202828322.pdf
week1slides1704202828322.pdf
 
Data Kinetics Products
Data Kinetics ProductsData Kinetics Products
Data Kinetics Products
 
Redshift overview
Redshift overviewRedshift overview
Redshift overview
 
Data Virtualization Reference Architectures: Correctly Architecting your Solu...
Data Virtualization Reference Architectures: Correctly Architecting your Solu...Data Virtualization Reference Architectures: Correctly Architecting your Solu...
Data Virtualization Reference Architectures: Correctly Architecting your Solu...
 
Still All on One Server: Perforce at Scale
Still All on One Server: Perforce at Scale Still All on One Server: Perforce at Scale
Still All on One Server: Perforce at Scale
 
Elastic storage in the cloud session 5224 final v2
Elastic storage in the cloud session 5224 final v2Elastic storage in the cloud session 5224 final v2
Elastic storage in the cloud session 5224 final v2
 
Gruter TECHDAY 2014 Realtime Processing in Telco
Gruter TECHDAY 2014 Realtime Processing in TelcoGruter TECHDAY 2014 Realtime Processing in Telco
Gruter TECHDAY 2014 Realtime Processing in Telco
 
Comparative study of modern databases
Comparative study of modern databasesComparative study of modern databases
Comparative study of modern databases
 
AWS Cloud experience concepts tips and tricks
AWS Cloud experience concepts tips and tricksAWS Cloud experience concepts tips and tricks
AWS Cloud experience concepts tips and tricks
 
S016828 storage-tiering-nola-v1710b
S016828 storage-tiering-nola-v1710bS016828 storage-tiering-nola-v1710b
S016828 storage-tiering-nola-v1710b
 
Best Practices – Extreme Performance with Data Warehousing on Oracle Databa...
Best Practices –  Extreme Performance with Data Warehousing  on Oracle Databa...Best Practices –  Extreme Performance with Data Warehousing  on Oracle Databa...
Best Practices – Extreme Performance with Data Warehousing on Oracle Databa...
 
A12 vercelletto indexing_techniques
A12 vercelletto indexing_techniquesA12 vercelletto indexing_techniques
A12 vercelletto indexing_techniques
 
Lecture 05 - The Data Warehouse and Technology
Lecture 05 - The Data Warehouse and TechnologyLecture 05 - The Data Warehouse and Technology
Lecture 05 - The Data Warehouse and Technology
 
Slides: Start Small, Grow Big with a Unified Scale-Out Infrastructure
Slides: Start Small, Grow Big with a Unified Scale-Out InfrastructureSlides: Start Small, Grow Big with a Unified Scale-Out Infrastructure
Slides: Start Small, Grow Big with a Unified Scale-Out Infrastructure
 

Recently uploaded

SWEBOK and Education at FUSE Okinawa 2024
SWEBOK and Education at FUSE Okinawa 2024SWEBOK and Education at FUSE Okinawa 2024
SWEBOK and Education at FUSE Okinawa 2024
Hironori Washizaki
 
LORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOM
LORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOMLORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOM
LORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOM
lorraineandreiamcidl
 
What is Augmented Reality Image Tracking
What is Augmented Reality Image TrackingWhat is Augmented Reality Image Tracking
What is Augmented Reality Image Tracking
pavan998932
 
GraphSummit Paris - The art of the possible with Graph Technology
GraphSummit Paris - The art of the possible with Graph TechnologyGraphSummit Paris - The art of the possible with Graph Technology
GraphSummit Paris - The art of the possible with Graph Technology
Neo4j
 
Artificia Intellicence and XPath Extension Functions
Artificia Intellicence and XPath Extension FunctionsArtificia Intellicence and XPath Extension Functions
Artificia Intellicence and XPath Extension Functions
Octavian Nadolu
 
2024 eCommerceDays Toulouse - Sylius 2.0.pdf
2024 eCommerceDays Toulouse - Sylius 2.0.pdf2024 eCommerceDays Toulouse - Sylius 2.0.pdf
2024 eCommerceDays Toulouse - Sylius 2.0.pdf
Łukasz Chruściel
 
Essentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FMEEssentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FME
Safe Software
 
Using Xen Hypervisor for Functional Safety
Using Xen Hypervisor for Functional SafetyUsing Xen Hypervisor for Functional Safety
Using Xen Hypervisor for Functional Safety
Ayan Halder
 
Automated software refactoring with OpenRewrite and Generative AI.pptx.pdf
Automated software refactoring with OpenRewrite and Generative AI.pptx.pdfAutomated software refactoring with OpenRewrite and Generative AI.pptx.pdf
Automated software refactoring with OpenRewrite and Generative AI.pptx.pdf
timtebeek1
 
openEuler Case Study - The Journey to Supply Chain Security
openEuler Case Study - The Journey to Supply Chain SecurityopenEuler Case Study - The Journey to Supply Chain Security
openEuler Case Study - The Journey to Supply Chain Security
Shane Coughlan
 
Graspan: A Big Data System for Big Code Analysis
Graspan: A Big Data System for Big Code AnalysisGraspan: A Big Data System for Big Code Analysis
Graspan: A Big Data System for Big Code Analysis
Aftab Hussain
 
AI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI App
AI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI AppAI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI App
AI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI App
Google
 
Vitthal Shirke Java Microservices Resume.pdf
Vitthal Shirke Java Microservices Resume.pdfVitthal Shirke Java Microservices Resume.pdf
Vitthal Shirke Java Microservices Resume.pdf
Vitthal Shirke
 
Atelier - Innover avec l’IA Générative et les graphes de connaissances
Atelier - Innover avec l’IA Générative et les graphes de connaissancesAtelier - Innover avec l’IA Générative et les graphes de connaissances
Atelier - Innover avec l’IA Générative et les graphes de connaissances
Neo4j
 
Empowering Growth with Best Software Development Company in Noida - Deuglo
Empowering Growth with Best Software  Development Company in Noida - DeugloEmpowering Growth with Best Software  Development Company in Noida - Deuglo
Empowering Growth with Best Software Development Company in Noida - Deuglo
Deuglo Infosystem Pvt Ltd
 
Orion Context Broker introduction 20240604
Orion Context Broker introduction 20240604Orion Context Broker introduction 20240604
Orion Context Broker introduction 20240604
Fermin Galan
 
Introducing Crescat - Event Management Software for Venues, Festivals and Eve...
Introducing Crescat - Event Management Software for Venues, Festivals and Eve...Introducing Crescat - Event Management Software for Venues, Festivals and Eve...
Introducing Crescat - Event Management Software for Venues, Festivals and Eve...
Crescat
 
GOING AOT WITH GRAALVM FOR SPRING BOOT (SPRING IO)
GOING AOT WITH GRAALVM FOR  SPRING BOOT (SPRING IO)GOING AOT WITH GRAALVM FOR  SPRING BOOT (SPRING IO)
GOING AOT WITH GRAALVM FOR SPRING BOOT (SPRING IO)
Alina Yurenko
 
Top Features to Include in Your Winzo Clone App for Business Growth (4).pptx
Top Features to Include in Your Winzo Clone App for Business Growth (4).pptxTop Features to Include in Your Winzo Clone App for Business Growth (4).pptx
Top Features to Include in Your Winzo Clone App for Business Growth (4).pptx
rickgrimesss22
 
Cracking the code review at SpringIO 2024
Cracking the code review at SpringIO 2024Cracking the code review at SpringIO 2024
Cracking the code review at SpringIO 2024
Paco van Beckhoven
 

Recently uploaded (20)

SWEBOK and Education at FUSE Okinawa 2024
SWEBOK and Education at FUSE Okinawa 2024SWEBOK and Education at FUSE Okinawa 2024
SWEBOK and Education at FUSE Okinawa 2024
 
LORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOM
LORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOMLORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOM
LORRAINE ANDREI_LEQUIGAN_HOW TO USE ZOOM
 
What is Augmented Reality Image Tracking
What is Augmented Reality Image TrackingWhat is Augmented Reality Image Tracking
What is Augmented Reality Image Tracking
 
GraphSummit Paris - The art of the possible with Graph Technology
GraphSummit Paris - The art of the possible with Graph TechnologyGraphSummit Paris - The art of the possible with Graph Technology
GraphSummit Paris - The art of the possible with Graph Technology
 
Artificia Intellicence and XPath Extension Functions
Artificia Intellicence and XPath Extension FunctionsArtificia Intellicence and XPath Extension Functions
Artificia Intellicence and XPath Extension Functions
 
2024 eCommerceDays Toulouse - Sylius 2.0.pdf
2024 eCommerceDays Toulouse - Sylius 2.0.pdf2024 eCommerceDays Toulouse - Sylius 2.0.pdf
2024 eCommerceDays Toulouse - Sylius 2.0.pdf
 
Essentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FMEEssentials of Automations: The Art of Triggers and Actions in FME
Essentials of Automations: The Art of Triggers and Actions in FME
 
Using Xen Hypervisor for Functional Safety
Using Xen Hypervisor for Functional SafetyUsing Xen Hypervisor for Functional Safety
Using Xen Hypervisor for Functional Safety
 
Automated software refactoring with OpenRewrite and Generative AI.pptx.pdf
Automated software refactoring with OpenRewrite and Generative AI.pptx.pdfAutomated software refactoring with OpenRewrite and Generative AI.pptx.pdf
Automated software refactoring with OpenRewrite and Generative AI.pptx.pdf
 
openEuler Case Study - The Journey to Supply Chain Security
openEuler Case Study - The Journey to Supply Chain SecurityopenEuler Case Study - The Journey to Supply Chain Security
openEuler Case Study - The Journey to Supply Chain Security
 
Graspan: A Big Data System for Big Code Analysis
Graspan: A Big Data System for Big Code AnalysisGraspan: A Big Data System for Big Code Analysis
Graspan: A Big Data System for Big Code Analysis
 
AI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI App
AI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI AppAI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI App
AI Fusion Buddy Review: Brand New, Groundbreaking Gemini-Powered AI App
 
Vitthal Shirke Java Microservices Resume.pdf
Vitthal Shirke Java Microservices Resume.pdfVitthal Shirke Java Microservices Resume.pdf
Vitthal Shirke Java Microservices Resume.pdf
 
Atelier - Innover avec l’IA Générative et les graphes de connaissances
Atelier - Innover avec l’IA Générative et les graphes de connaissancesAtelier - Innover avec l’IA Générative et les graphes de connaissances
Atelier - Innover avec l’IA Générative et les graphes de connaissances
 
Empowering Growth with Best Software Development Company in Noida - Deuglo
Empowering Growth with Best Software  Development Company in Noida - DeugloEmpowering Growth with Best Software  Development Company in Noida - Deuglo
Empowering Growth with Best Software Development Company in Noida - Deuglo
 
Orion Context Broker introduction 20240604
Orion Context Broker introduction 20240604Orion Context Broker introduction 20240604
Orion Context Broker introduction 20240604
 
Introducing Crescat - Event Management Software for Venues, Festivals and Eve...
Introducing Crescat - Event Management Software for Venues, Festivals and Eve...Introducing Crescat - Event Management Software for Venues, Festivals and Eve...
Introducing Crescat - Event Management Software for Venues, Festivals and Eve...
 
GOING AOT WITH GRAALVM FOR SPRING BOOT (SPRING IO)
GOING AOT WITH GRAALVM FOR  SPRING BOOT (SPRING IO)GOING AOT WITH GRAALVM FOR  SPRING BOOT (SPRING IO)
GOING AOT WITH GRAALVM FOR SPRING BOOT (SPRING IO)
 
Top Features to Include in Your Winzo Clone App for Business Growth (4).pptx
Top Features to Include in Your Winzo Clone App for Business Growth (4).pptxTop Features to Include in Your Winzo Clone App for Business Growth (4).pptx
Top Features to Include in Your Winzo Clone App for Business Growth (4).pptx
 
Cracking the code review at SpringIO 2024
Cracking the code review at SpringIO 2024Cracking the code review at SpringIO 2024
Cracking the code review at SpringIO 2024
 

InfiniFlux vs_RDBMS

  • 2. Overview 2 • The concept of InfiniFlux: the ultra high-speed database that stores and processes time series data. • InfiniFlux has very different characteristics compared to the conventional DBMS such as Oracle and DB2 in order to provide high-speed processing. • Importance of understanding: the characteristics and architecture for time series log data. • Technical characteristics: analyzing data using SQL and storing hundreds of thousands of records in a second in real-time. • Describe the differences between InfiniFlux and the conventional technology, and provide detailed explanation about every item. Document Overview
  • 3. Comparison Chart 3 Characteristics InfiniFlux DB RDBMS Transaction Provide implicit transaction based on Snapshot Provide explicit transaction based on log file INSERT speed 300,000 ~ 3,000,000 in a second less than 10,000 in a second SELECT performance Optimized for query over search & statistical analysis Optimized for OLTP Characteristics of Data Time series log data Non-log transaction data Updatable data Append Only (Write Once, Read Many) Updatable DELETE operation Delete the oldest data Delete random data Real-time index Real-time bitmap index Not real-time B+Tree Data compression High performance data compression in real- time Not support data compression Real-time search for text Supported (real-time inverted index) Not supported (Even if supported, not in real-time) Support time series data Support data partitioning by sharding Support data partitioning based on general timestamp column Gap between data and index Occurrence of momentary gap Gap not occurred
  • 4. Support Transaction 4 RDBMS • Provide explicit transaction for all the data operation • Transaction: set of operations that satisfy ACID properties • Atomicity • Consistency • Isolation • Durability • Savepoint, Commit, and Rollback InfiniFlux • Not provide explicit transaction • No transaction for data operation (input) • Provide implicit transaction over internal meta data • Table structure, index structure, and data file structure Reasons for not supporting transaction • No need to store data based on transaction since it is time series data • Fast storage and processing are much more valuable than transaction • No need to pay costs and conduct complex operations for transactions
  • 5. Input Performance 5 RDBMS • Difficult to input data more than 5,000 per second • High costs of logging for transaction. As a result, increase I/O costs. • High costs of index update. Thus, B+Tree is suitable for data search. • To maintain the consistency, all the operations of index update for a record is conducted in consecutive order. • Degrading system performance as the volume of index is greater than data InfiniFlux • Able to input hundreds of thousands data per second • Efficient to create index through real-time bitmap index • Costs for logging not required • Parallel index can be created using multi-threads • Reduce the amount of I/O data by real-time compression. As a result, improve overall system performance. • Able to improve performance greatly by creating tablespace based on multiple disks
  • 6. Query Performance 6 RDBMS • Row-oriented database has advantage over online transaction processing (OLTP). • High query performance on high cardinality • Reason: small search range brings high performance like B+Tree • Mainly operating based on B+Tree, and select the most efficient index and use it. • Efficient to search in a small range even with large number of data • Slow query performance for statistical analysis InfiniFlux • Column-oriented database has advantage over online analytical processing (OLAP). • High query performance on low cardinality • Most of time series log data are low cardinality since it has high level of duplication. • Most of them are bitmap indices, and it is efficient since more than two indices can be used at the same time. • Fast statistical query against massive data (hundreds of millions of data) • Relatively slow on search for a certain record of OLTP from the whole DB • In this case, global index needs to be created.
  • 7. Characteristics of Data 7 RDBMS • Optimized for storing data through transactions • Financial information and individual identification information for banking transactions • Important data that should be safely stored in conventional database such as Oracle. • RDMBS is not related with the flow of time and can be updated or deleted. InfiniFlux • Target data are time series log data. • Hundreds of thousands of data were created in a second. • Update operation is not required and shows high level of data duplication. • Target data are log files or similar data that were previously stored as text files. • Constantly describe the status of a certain target over time.
  • 8. Updatable Data 8 RDBMS • Updatable model • Data can be updated or modified anytime. • Able to delete an arbitrary record at a random moment. • Create a database that can access and modify all the data with ease. InfiniFlux • Write Once Read Many (WORM) model • Since it is time series log data, data cannot be modified once it is stored. • Read-oriented, and not support UPDATE operation at all • Deleting data? • Cannot delete a record at a certain time • Able to delete the oldest record from the database in sequential order * InfiniFlux provides delete feature in order to maintain a certain level of disk usage of embedded device.
  • 9. Index Technology 9 RDBMS • In general, it uses B+Tree • It is global index and all the record information are stored in the index. • Expensive costs of logging and recovery operations for supporting transactions * Due to the reasons mentioned above, it process only thousands data per second. • Disk usage is increasing rapidly compared to the number of indices because values of raw data are stored in the index. • Not provide compression feature due to performance issue InfiniFlux • Support real-time bitmap index • Local index structure of partition unit • Composed with various indices over all the records • Able to create index quickly since transaction support and logging operations are not required. • Able to create millions of indices per second • Costs of creating index will increase little by little even though the amount of data increased. • Minimize the disk usage by supporting compression algorithms • OLTP query is relatively slow when the amount of data increases indefinitely. • Support semi-global index (4th quarter 2015) • Able to provide high performance on OLTP query • Able to provide high performance on compression and statistical query
  • 10. Compression Technology 10 RDBMS • No data compression issues. * The reason for this is that, in general, the number of target records are maintained at a certain number. In this environment, disk usage is not that big issue. • The basic idea is that increase search performance at the expense of disk space. • In many cases, clients are not interested in data compression itself. • Even if compression is available, challenges below will be presented; • Increase the amount of data usage due to the structure in data duplication of B+Tree • Difficult to use the duplication property of row-oriented database InfiniFlux • Amount of disk usage is very important since it is assumed that the environment where large amount of data is created. • Conduct real-time compression twice and thus able to use disk space efficiently • Logical compression • Compress column-based duplicated data in a dictionary structure • Duplicated bit strings over bitmap index at high compression logically • Physical compression • Real-time physical compression of partition pages when storing disks • Disk usages show slow growth pattern not linear growth as the number of index increases.
  • 11. Full Text Search 11 RDBMS • In general, database do not support text search • Provide alternative methods such as LIKE statement for the search. • Partial search over the column were conducted through LIKE ‘%pattern%’. In this case, full scan over all the records is operated, and resulted in slow performance. • Impossible to use RDBMS for full text search InfiniFlux • Provide keyword index, and full text search is available for a certain VARCHAR column. • “SEARCH” can be used for searching instead of “LIKE”. • SELECT id FROM user_table WHERE address SEARCH ‘Texas’; • Able to search text in UTF-8 format. • Different processing methods are applied to English characters and 2 byte character (Chinese/Japanese/Korean). • English • Separate words with special characters (e.g. “boy” and “boys” are different) • CJK (Chinese, Japanese, and Korean) • Using 2-gram method to search • The word ‘강남구’ will be converted into ‘강남’ & ‘남구’ and operate
  • 12. Concurrency Level 12 RDBMS • Generally provide “Record Level Locking” • Two options based on how to read records while updating • Consistency Read • Allow to read previous value of the record that is being updated. • No confliction of lock  No waiting time • Non Consistency Read (Record locking conflict occurred) • Allow to wait and decide to commit/rollback on the record that is being updated. • Conflicts arise while locking  Possible to wait indefinitely depends on the previous transaction. InfiniFlux • Provide Lockless structure • No locking conflict due to no update • No lock mechanism against the records, thus, maximize the performance over data search and analysis.
  • 13. Time Series Data Analysis 13 Sliding Memory Window Memory Window Memory Window File -1 File -2 File -3 File -4 File -5 File -6 File -7 Data Insert Current time Old time RDBMS • Time series data exist in the form of time data type (date, time, timestamp, and interval). • It doesn’t recognize the data as time series data, rather treat it with other general data (number, and varchar etc.). • Separate devices for analyzing data based on the time base is not provided due to the reasons above. • Slow performance when analyzing data with time index. InfiniFlux • When data were entered, it creates physical partitions in sequential order. • Able to access records in a certain time range directly due to the sequential order. • When data were entered, the timestamp for the record will be stored in nano unit automatically (_arrival_time as a hidden column). • Able to operate data of the time base freely based on the hidden column. • Select data from table where DURATION 10 minute (output data within 10 minutes)
  • 14. Backup and Restore 14 Backup Restore RDBMS Backup data in a certain table or database area externally • Online backup is the basic strategy to store data • Separately provide incremental backup in order to reduce the time and costs of backup The process of returning the data to the original database • Impossible to use backup files and must go through restore process • Backup big files take a long time to restore • Able to backup data based on a certain time, but complicate to change the starting point. InfiniFlux • Backup the whole database based on a certain time • Not able to backup data based on a certain table or record • Backup in a file or directory format multiple files • Restore the whole backup files • Overwrite the existing database
  • 15. Backup and Mount 15 What is mount? • Rather than storing backup database file, it loads and searches the contents of database after loading only meta data. • Similar to the concept of the disk mount of UNIX • Able to access backup data directly in a certain time in a fast and efficient way. RDBMS • Not supported InfiniFlux • MOUNT • Mount backup data on 31st December 2014 e.g.) MOUNT DATABASE ‘/home/data/2014-12-31’ • UNMOUNT • Unmount the mounted data e.g.) UNMOUNT DATABASE ‘/home/data/2014-12-31’
  • 16. The World's Fastest Time Series DBMS for IoT and Big Data www.infiniflux.com info@infiniflux.com InfiniFlux