NoSQL – Data Center Centric Application Enablement

879 views
663 views

Published on

The growth of Datacenter infrastructure is trending out of bounds, along with the pace in user activity and data generation in this digital era. However, the nature of the typical application deployment within the data center is changing to accommodate new business needs. Those changes introduce complexities in application deployment architecture and design, which cascade into requirements for a new generation of database technology (NoSQL) destined to ease that complexity. This webcast will discuss the modern data centers data centric application, the complexities that must be dealt with and common architectures found to describe and prescribe new data center aware services. Well look at the practical issues in implementation and overview current state of art in NoSQL database technology solving the problems of data center awareness in application development.

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
879
On SlideShare
0
From Embeds
0
Number of Embeds
95
Actions
Shares
0
Downloads
20
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

NoSQL – Data Center Centric Application Enablement

  1. 1. Data Centers and NoSQL Database Robert Greene – Product Management
  2. 2.  Data Center Trends and Business Drivers  Key Tech Requirements  The Data Center Aware Application  3 NoSQL Data Center Architectures  Data Locality and Reliability Isolation  Data Consistency Considerations  NoSQL Data Center Implementation  Summary - NoSQL & Data Centers NoSQL - Enabling Data Center Applications
  3. 3. Data Centers - Trends and Drivers  Trends – More deployments • SMB’s going global • Consumer facing • Terabytes of Data • Proximity and Rev • Requirements • Security • Regulatory • Availability  IDC survey revenue>$1B or employees>5000 • 26% have 6 or more DC’s
  4. 4.  Cloud deployments • SMB’s going global • Consumer facing • Terabytes of Data • Proximity and Rev • Requirements • Security • Regulatory • Availability Data Centers - Trends and Drivers
  5. 5.  Business Drivers • Cost Optimization • SaaS and Desktop Virtualization • Globalized Sensor Data and M2M • Big Data – Physics of IT • Revenue • Latency and Availability Data Centers - Trends and Drivers
  6. 6.  Latency reduction • Amazon study • 100ms = 1% Revenue • 150M dollars • SMB seeing more data • Physics • More data = more time, • Longer distance = more time Key Technical Requirements
  7. 7.  Availability • 2003 New York Blackout • Dec 2012 AWS Outage • Regulatory Mandates Key Technical Requirements
  8. 8.  Process & Data Redundancy • Specialized Needs • Provisioning • Placement • Communications • Encryption • Security • Data Availability • Upgrade processing • Failover processes The Data Center Aware Application
  9. 9.  3 Prevalent Architectures • Container • Component • Tag NoSQL Data Center Architectures Data Replica Replica Data Center 1 Data Center 2 Data Center 3
  10. 10. NoSQL Data Center Deployment Architectures Node Data Replica Replica Data Replica Replica Hardware Unit of Replication Process Container Based  Strengths • Allows Local Writes • Simple Replication Admin • Good Global Availability  Weakness • Many Data Copies • High Bandwidth Rep • No Consistent Reads Durability & ConsistencyDurability & Consistency Replication More Sites
  11. 11. NoSQL Data Center Deployment Architectures Data Replica Replica Hardware Unit of Replication Process Component Based  Strengths • Minimal Data Copies • Low Bandwidth Rep • Good Regional Availability • Simple Admin  Weakness • Network Latency Sensitive • Data placement Durability & ConsistencySync Channel
  12. 12.  Strengths • Complete Control • Targeted Data NoSQL Data Center Deployment Architectures Node Data Replica Replica Data Replica Replica Hardware Unit of Replication Process Data Replica Replica TAG_A TAG_ATAG_A Tag Based  Weakness • Coding Specific • Brittle to system change • Complex management • Non-optimal data copies Durability & Consistency Durability & Consistency Sync Channel
  13. 13.  Hybrid Architectures • Component (local reliability) • Container (multi channel)  Cloud Architectures • Model Physical Infrastructure • Geographic Regions • Zoned local isolation points • Power & Network only NoSQL Data Center Architectures
  14. 14.  Physical - Points of Failure Isolation • Disk, Server, Rack, Power, Network Switches  Latency Placement Considerations Data Locality and Reliability Isolation Region Data Center Zone 3Zone 2Zone 1 RD R DR R A Net Switches Servers Disks High/Low Latency Racks Net Switches Net Switches Low Latency Power
  15. 15.  Multiple locations need to be kept in sync  Latency – Read Consistency across Data Centers • Consistency based on eventually consistent processes • How to ensure you have the latest data if needed • How to read locally to keep latency low  Throughput – Write Durability across Data Centers • Durability based on write copies, eventually everywhere • How to get copies written without high latency • How to resolve conflicting updates  Techniques: quorum voting, vector clocks, timestamp Data Consistency Considerations
  16. 16.  Container Architecture • Reads and writes always local to the container • No option to ensure consistent data • Sync involves key differentiation scheme ( e.g. Merkle tree)  Component Architecture • Reads and writes optionally to local component • Can request consistent operation ( possible latency cost ) • Sync involves log ordered sequencing ( no differentiation )  Tagged Architecture • Reads and writes explicit at the data level • Can request consistent operation ( possible latency cost ) • Sync is application code dependent Data Consistency Considerations
  17. 17.  Review of a few well known vendors • Cassandra • Hbase • MongoDB • OnDB • Riak NoSQL Data Center Implementation
  18. 18. Cassandra Data Center Implementation  Data Center Architecture: Container  Replication Unit: Node  Reliability: Cluster  Data Copies: RF * Data Center  Durability: One, Local_Quorum  Consistency: Timestamp  Placement: Multiple copies per Data Center
  19. 19. Hbase Data Center Implementation  Data Center Architecture: Tag  Replication Unit: Column Family  Reliability: Cluster  Data Copies: RF x Clusters (family subset)  Durability: ACID  Consistency: Absolute  Placement: None  Extensions: • Multi-Channel Replication ( per region server )
  20. 20. MongoDB Data Center Implementation  Data Center Architecture: Tag  Replication Unit: Range of Collection  Reliability: Replica Set  Data Copies: RF x Tagged Replica Set (range subset)  Durability: WriteConcern  Consistency: ReadConcern  Placement: Hard Coded  Extensions: • Read-Only Shards
  21. 21. Riak Data Center Implementation  Data Center Architecture: Container  Replication Unit: Cluster  Reliability: Local Cluster  Data Copies: RF x Clusters  Durability: One, Quorum, All  Consistency: Vector Clock  Placement: Multiple copies per cluster  Extensions: • Multi-Channel Sink • Read-Only Cluster
  22. 22.  Data Center Architecture: Component  Replication Unit: Zone  Reliability: Zone(s)  Data Copies: RF  Durability: ACID, One, Quorum, All  Consistency: Absolute, Quorum  Placement: Copy per Zone  Extensions: • Local Quorum • Read-Only Zones OnDB – Oracle NoSQL DB Data Center Data Replica Replica Zone Near Data Centers Client ReadClient Write Availability Group
  23. 23.  Data Center Deployments • Increasingly Common: Cost, Revenue & Reliability Advantages  The Data Center Aware Application ( latency & consistency ) Summary – NoSQL & Data Centers Component X Container X X Tag X X Near Data Centers X X X Multi Channel X Disaster Recovery Placement Far Data Centers Placement X X Placement Consistent Data X X In-Consistent Data X X X Data Size - Copies Low High High Low Medium
  24. 24. Oracle NoSQL DB Resources • NoSQL Database Downloads http://www.oracle.com/technetwork/products/nosqldb/downloads/index.html • NoSQL Database Documentation http://www.oracle.com/technetwork/products/nosqldb/documentation/index.html • NoSQL Database Contacts • David Segleau – Director Product Management dave.segleau@oracle.com • Robert Greene – Senior Principle Product Manager robert.c.greene@oracle.com

×