+ All Categories
Home > Technology > Thesis finalpresentation

Thesis finalpresentation

Date post: 28-Nov-2014
Category:
Upload: julia-proskurnia
View: 1,815 times
Download: 1 times
Share this document with a friend
Description:
In recent years the need for distributed data storage has led the way to design new systems in a large-scale environment. The growth of unbounded stream of data, the necessity to store and analyze it in real time, reliably, scalably and fast are the reasons for appearance of such systems in financial sector, stock exchange Nasdaq OMX especially. Futhermore, internally designed totally ordered reliable message bus is used in Nasdaq OMX for almost all internal subsystems. Theoretical and practical extensive studies on reliable totally ordered multicast were made in academia and it was proven to serve as a fundamental block in construction of distributed fault-tolerant applications. In this work, we are leveraging NOMX low-latency reliable totally ordered message bus with a capacity of at least 2 million messages per second to build high performance distributed data store. The data operations consistency can be easily achieved by using the messaging bus as it forwards all messages in reliable total order fashion. Moreover, relying on the reliable totally ordered messaging, active in-memory replication support for fault tolerance and load balancing is integrated. Consequently, the prototype was developed using pro- duction environment requirements to demonstrate its feasibility. Experimental results show a great scalability, and performace serving around 400,000 insert operations per second over 6 data nodes that can be served with 100 microseconds latency. Latency for single record read operations are bound to sub-half millisecond, while data ranges are retrieved with sub-100 Mbps ca- pacity from one node. Moreover, performance improvements under a greater number of data store nodes are shown for both writes and reads. It is con- cluded that uniform totally ordered sequenced input data can be used in real time for large-scale distributed data storage to maintain strong consistency, fault-tolernace and high performance.
29
GDS: Genium Data Store Real Time, Low Latency, Reliable Iuliia Proskurnia EMDC KTH 2013
Transcript
Page 1: Thesis finalpresentation

GDS: Genium Data Store Real Time, Low Latency, Reliable!

Iuliia Proskurnia!EMDC!KTH!2013!

Page 2: Thesis finalpresentation

2!

Page 3: Thesis finalpresentation

3!

Page 4: Thesis finalpresentation

4!

3900 companies! 39 countries! over 1500 corporate products!

USE CASE!

Write events!Retrieve ranges of records!

Page 5: Thesis finalpresentation

5!

Fault-Tolerant?!

Consistent?!

Fast?!

Scalable?!

Page 6: Thesis finalpresentation

6!

Approaches

!   Consensus based!

!   ...!!   Total Order Multicast!

!   Symmetric!

!   Token Site !

Page 7: Thesis finalpresentation

Uniform Reliable Total Order

◦  Validity !

◦  Uniform Integrity !

◦  Uniform Agreement !

◦  Uniform Total Order !

!

7!

Page 8: Thesis finalpresentation

8!

Genium INET Message Bus �Uniform Reliable Total Order Multicast

!   Similar to Amoeba protocol!

!  However... Fault Tolerant!!

Page 9: Thesis finalpresentation

9!

GDS: Genium Data Store

!   Uses Genium INET Message Bus abstraction!

!   Clients, Sequencer, Data store!

! Rewinders and sequencer replication!

!   Active replication!

Client!

Data store node!

Data store node!

Page 10: Thesis finalpresentation

10!

GDS high level abstraction

LEDS!

Page 11: Thesis finalpresentation

11!

LEDS

!   Column based!

!   BLOBS!

!   Appends!

!   Range Queries!

!  Not Distributed!

!  Not fault-tolerant!

Page 12: Thesis finalpresentation

12!

Properties

!   Consistent!

!   Failure Resilient!!   Replication!

!   Rewinders!

!   Cite Replication!

Total Order!

Page 13: Thesis finalpresentation

13!

Possible Failure Scenarios C

lient

Fai

lure!

Sequ

ence

r Fai

lure!

���8VHU

6HT

'6

���

>Q@

>Q@

:RUN

>Q��@

&XUUHQW�0HVVDJH��>Q��@ 5HZLQG�0HVVDJH��>Q@

���8VHU

6HT

'6

��� 5HZLQG�0HVVDJH��>Q@

6HT6WDQG%\

>Q@

>Q@

>Q@

>Q��@

3ULPDU\�3�

���

3

>Q@

>Q@

>Q@

5HZLQG�0HVVDJH��>Q@

Page 14: Thesis finalpresentation

14!

Scalability

!  Natural Load Balancing!

!   Partitioning (manual)!

Page 15: Thesis finalpresentation

15!

Evaluation

!   Inserts (throughput/latency)!

!   Range Queries (throughput)!

!   Range transmission failure!

Page 16: Thesis finalpresentation

16!

Set Up

Page 17: Thesis finalpresentation

17!

Writes �Throughput

Page 18: Thesis finalpresentation

18!

Writes Limits

Page 19: Thesis finalpresentation

19!

Writes �Latency

Page 20: Thesis finalpresentation

20!

Range Queries �Throughput

Page 21: Thesis finalpresentation

21!

Range Queries �Scalability

8 Concurrent Users!

Page 22: Thesis finalpresentation

22!

Range Queries �Link Failure

Page 23: Thesis finalpresentation

23!

Summary

!   uniform reliable total order multicast !

!   scales fine!

!   low latency!

!   consistent, fault-tolerant!

Page 24: Thesis finalpresentation

24!

Future Work

!   Generality!

!   Send compressed chunks!

!   Automated partitioning!

!   Long-running tests!

Page 25: Thesis finalpresentation

25!

Comments? Questions?

Thesis Writing Process!

Page 26: Thesis finalpresentation

26!

Single record read�without load

Page 27: Thesis finalpresentation

27!

Single record read�with load (10 000 inserts)

Page 28: Thesis finalpresentation

28!

Single record read�scalability

Page 29: Thesis finalpresentation

29!

Discussion


Recommended