Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Strata London WANdisco

44 views

Published on

Companies that understand how to overcome Data Gravity with a LiveData strategy stand to benefit the most from innovations in data and analytics. Others will continue to invest in legacy systems and data silos that fail to take advantage of distributed computing and the rapid pace of data product innovation on the cloud. This talk shares best practices our clients have taken to evolve their data architecture to become a LiveData company.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Strata London WANdisco

  1. 1. © 2019 WANdisco PLC. All rights reserved. Overcome Data Gravity with A LiveData Strategy Strata Data 2019 — London Joel S. Horwitz, Senior VP Marketing, WANdisco
  2. 2. © 2019 WANdisco PLC. All rights reserved.
  3. 3. © 2019 WANdisco PLC. All rights reserved.
  4. 4. © 2019 WANdisco PLC. All rights reserved.
  5. 5. © 2019 WANdisco PLC. All rights reserved. Wide Area Network Distributed Computing The experts in distributed computing A decentralized IT organization requires a Platform to ensure information availability and consistency across multiple data centers, development sites, data clusters, and multiple clouds. © 2019 WANdisco PLC. All rights reserved.
  6. 6. © 2019 WANdisco PLC. All rights reserved. A platform for ANY IT Architecture MultiSite Active Backup Disaster Recovery MultiCloud Data Lake Migration FUSION Security DConE Web UI SDK LiveCode Git Gerritt SVN LiveData Local/NFS HDFS Object Storage IBM Alibaba Cloud AWS Azure Google Cloud Oracle Cloud Architect Security Manager Data Architect Cloud Engineer Data Operator Data Engineer Persona Solutions Platform IT Assets Cloud
  7. 7. © 2019 WANdisco PLC. All rights reserved. WANdisco FUSION Platform Protect Your Investment No downtime, no outages, and no risk with guaranteed near- zero RTO and RPO. Transform Your IT Economics Create a bedrock for performance by fully utilizing hardware previously reserved for backup and recovery. Break through constraints Distribute all of your code and data for developers to access at the speed of business.
  8. 8. © 2019 WANdisco PLC. All rights reserved. Hadoop to Hadoop
  9. 9. © 2019 WANdisco PLC. All rights reserved.
  10. 10. © 2019 WANdisco PLC. All rights reserved. Customer Example: Hadoop to Hadoop “One of the things that impressed us most about the WANdisco Fusion platform was its ability to resolve the data consistency challenges of replicating large amounts of data between active Hadoop clusters.” Steven Swartzlander, Lead Architect at HM Health Solutions
  11. 11. © 2019 WANdisco PLC. All rights reserved. Hadoop to Cloud
  12. 12. © 2019 WANdisco PLC. All rights reserved. Customer Example: Hadoop to Cloud Playtika used WANdisco Fusion to transfer data to Amazon S3 as it changed in their on-premises Cloudera Hadoop cluster. WANdisco Fusion’s active data replication guaranteed data consistency in both locations.
  13. 13. © 2019 WANdisco PLC. All rights reserved. Cloud to Cloud
  14. 14. © 2019 WANdisco PLC. All rights reserved. Customer Example: Cloud to Cloud With WANdisco products, AMCO can replicate all incoming Application data from their hosted instance in Mexico to AWS.
  15. 15. © 2019 WANdisco PLC. All rights reserved. The 3 stages of Data Movement to Cloud BigData Movement to the Cloud Hybrid Data to Ensure Business Continuity LiveData for MultiCloud PHASE PHASE PHASE 321 • Single pass migration with guaranteed consistency • Minimal disruption to operations • Maximize target ingest rate • Every path becomes live as soon as it is migrated • Eliminate vendor lock-in • Gain true data availability • Reduce cost structures • Easy Management of Multiple Cloud Storage Providers • No Vendor Lock-in • Increased Data Availability
  16. 16. “Computer science still says what we have done is impossible… but it’s not.” Dr. Yeturu Aahlad, Chief Scientist, Inventor & Co-founder
  17. 17. © 2019 WANdisco PLC. All rights reserved. The LiveData Gap
  18. 18. © 2019 WANdisco PLC. All rights reserved. WANdisco eliminates the LiveData gap inconsistency, unused capacity, economic limitations Multiplies IT’s ability to execute at any scale, with same budget The LiveData Gap
  19. 19. © 2019 WANdisco PLC. All rights reserved. WANdisco eliminates the LiveData gap WANdisco Fusion multiplier effect: IT achieves more with the same budget Multiplies IT’s ability to execute at any scale, with same budget LiveData Multiplier Effect
  20. 20. © 2019 WANdisco PLC. All rights reserved. The LiveData Gap 20ZB global data storage in 2016, growing to 160Zb in 2025 $46B global storage spend in 2016 $67B global spend on public cloud computing Massive growth, massive investments: can’t keep up
  21. 21. © 2019 WANdisco PLC. All rights reserved. A platform for ANY IT Architecture
  22. 22. © 2019 WANdisco PLC. All rights reserved. Industries and Professions Automotive Government Healthcare Financial Services Telecom Energy Entertainment Retail Developer Collaboration Manufacturing
  23. 23. © 2019 WANdisco PLC. All rights reserved. “WANdisco Fusion provides consistent data across our cloud and on-premises solutions, offering near- zero RPO and enabling hybrid cloud agility to drive the business forward.” — Ajay Prasad, Big Data Leader, AMD
  24. 24. © 2019 WANdisco PLC. All rights reserved. “Of all the technologies we evaluated, WANdisco Fusion was the only solution that offered the speed, guaranteed data consistency, and continuous replication we required to analyze our patient data.” — Dr. Susheel Varma, Technology Officer for the Center for Computational Imaging & Simulation Technologies in Biomedicine, The University of Sheffield
  25. 25. © 2019 WANdisco PLC. All rights reserved. “WANdisco Fusion ensures that all of our sites are up and running all of the time.” — Angela Thomas, Development Tools Manager, Juniper Networks
  26. 26. © 2019 WANdisco PLC. All rights reserved. “WANdisco enabled us to upgrade our system effortlessly with no issues after we went live. We are a very happy and satisfied customer.” — Adinarayana Sreenath, Director, Big Data Architecture, Envestnet | Yodlee
  27. 27. © 2019 WANdisco PLC. All rights reserved. Our Strategic Partners WANdisco’s industry-leading technology and services partners are ready to help you deploy WANdisco Fusion and achieve LiveData.
  28. 28. © 2019 WANdisco PLC. All rights reserved. Forward Thinking Executives have a LiveData Strategy Fortune 50 Motor Vehicle Company 100% uptime while moving massive continuously changing sets of vehicular data for driverless car initiative Fortune 50 Retailer 2X compute capacity fully deployed by reducing server hardware requirements by 50% Global 50 Bank ZERO planned outages while seamlessly upgrading key financial applications
  29. 29. David Richards, CEO & founder, WANdisco “Founding a company is a personal & intimate process that demands huge amounts of passion, commitment & energy.”
  30. 30. © 2019 WANdisco PLC. All rights reserved. Demonstration
  31. 31. © 2019 WANdisco PLC. All rights reserved. WANdisco Fusion coordinates change
  32. 32. © 2019 WANdisco PLC. All rights reserved. Let’s follow a simple scenario of one object PUT performed by one client, using the LiveData for Multicloud Fusion product Object x A B C A B C
  33. 33. © 2019 WANdisco PLC. All rights reserved. User sends a request to create an object in object storage. This goes through the Fusion Multicloud Proxy Object x Proposal to create an object PUT A B C A B C
  34. 34. © 2019 WANdisco PLC. All rights reserved. Fusion Server gets this request from the Multicloud Proxy, and co- ordinates the request with other Fusion Servers in a membership Object x Consensus to create the object PUT A B C A B C
  35. 35. © 2019 WANdisco PLC. All rights reserved. This membership emits a Global Sequence of operations Object x Replica Agreements PUT A B C A B C
  36. 36. © 2019 WANdisco PLC. All rights reserved. Execution of the same Globally-Sequenced operation in each location results in replica object creation Object x PUT A B C A B C
  37. 37. © 2019 WANdisco PLC. All rights reserved. Originating client PUTs via the proxy to its local object storage Object x Object write PUT A B C A B C
  38. 38. © 2019 WANdisco PLC. All rights reserved. Object creation occurs locally at local speed with only the consensus performed in addition so far Object x A B C A B C
  39. 39. © 2019 WANdisco PLC. All rights reserved. Fusion is aware of the completion of that PUT and initiates the transfer of content to other replicas Object x Object x Content replication A B C A B C
  40. 40. © 2019 WANdisco PLC. All rights reserved. Writes can occur in any location, while Fusion guarantees the consistency of the object storage through consensus Objec t x Object x A B C A B C
  41. 41. © 2019 WANdisco PLC. All rights reserved. Multiple clients can operate in any location without changing that consistency guarantee
  42. 42. © 2019 WANdisco PLC. All rights reserved. Different types of underlying object storage can be used
  43. 43. © 2019 WANdisco PLC. All rights reserved. This can extend beyond what the public cloud vendors offer as internal, eventually-consistent replication only
  44. 44. © 2019 WANdisco PLC. All rights reserved. Fusion can operate across arbitrary distance
  45. 45. © 2019 WANdisco PLC. All rights reserved. THANK YOU Connect @wandisco
  46. 46. © 2019 WANdisco PLC. All rights reserved. Backup Slides
  47. 47. © 2019 WANdisco PLC. All rights reserved. Public Cloud A West Region FUSION enables LiveData everywhere Public Cloud A East Region Data in Production Dev/Test or Disaster Recovery Data Users Public Cloud B HYPERSCALE ECONOMICS Fully automated, utilizes all capacity, data at any scale ALWAYS ACCURATE Query any node, get the same results RUNS ANYWHERE On premises, hybrid cloud, cloud, multi-cloud ALWAYS PROTECTED No single point of failure ALWAYS AVAILABLE Shared logical pool of data
  48. 48. © 2019 WANdisco PLC. All rights reserved. WANdisco Fusion for Disaster Recovery WANdisco Fusion guarantees global data consistency and accessibility for your critical applications, and protects your data in the event of a disaster.
  49. 49. © 2019 WANdisco PLC. All rights reserved. WANdisco Fusion for Cloud MigrationWANdisco Fusion gives you the ability to successfully move to the cloud without risking business continuity.
  50. 50. © 2019 WANdisco PLC. All rights reserved. WANdisco Fusion for Hybrid Cloud Transition from separate management of data in private and public clouds to unified, integrated operations, easily and cost effectively.
  51. 51. © 2019 WANdisco PLC. All rights reserved. WANdisco Fusion for Multi-Cloud Replication Deploy WANdisco Fusion to enable global data access across any cloud provider so that you can access your data when you want it from anywhere.
  52. 52. © 2019 WANdisco PLC. All rights reserved. Approaches to data replication: DistCp, AzCopy, etc. • Unidirectional replication with a source and target • Target Storage/location is typically read only • Content will become inconsistent • Manual and labor intensive process to manage and reconcile differences • Inefficient use of the network • Only works between similar storage technologies • Does not cater for consistency of all relevant metadata (security, etc.) Step 2 Data from source is periodically copied into target Data Center 2 Target Storage Data Center 1 Source Storage ACTIVE NAMENODE DATA NODES ACTIVE NAMENODE DATA NODES INTERNET Step 1 Data is ingested into source first VPN POINT-IN-TIME or SCHEDULED COPY
  53. 53. © 2019 WANdisco PLC. All rights reserved. Approaches to data replication: • Disruptions in any of the separate environments will cause the two storage environments to diverge • Potential to run out of buffer at point of ingest when inter-environment network is down • Requires constant attention and administration to keep running • Data modified on any environment after ingest is not replicated, leading to further divergence INTERNET Data Center 2 Environment B Data Center 1 Environment A INTERNET Step 2 Load balancer sends one copy of data to Environment A and one copy to Environment B Step 1 Data is sent to load balancer LOAD BALANCER ACTIVE NAMENODE DATA NODES ACTIVE NAMENODE DATA NODES PARALLEL (DUAL) INGEST
  54. 54. © 2019 WANdisco PLC. All rights reserved. Approaches to data replication: • Global consistency not possible in a multi-cloud environment (vendor specific only, and limited to specific paired regions) • Primarily uni-directional, relies on versioning to avoid demands of true active/active operation • In some environments, replication is eventually consistent, and not guaranteed to be immediately available Cloud Provider A Region 1 Cloud Provider A Region 2 One-way replication Eventual consistency Cloud Provider A S3 api Cloud Provider B Other api Vendor specific APIs extract and load strategy CLOUD-NATIVE GEO REPLICATION & CDC
  55. 55. © 2019 WANdisco PLC. All rights reserved. WANdisco LiveData for MultiCloud LiveData for MultiCloud • Built on the WANdisco Fusion platform • Extends capabilities for object store consistency and replication to multi- cloud architectures • Spans S3-compatible, Azure Blob storage, on-premises and cloud storage systems • Same scale, performance, flexibility that WANdisco brings to big data platforms
  56. 56. © 2019 WANdisco PLC. All rights reserved. WANdisco LiveData for MultiCloud Capabilities • Data consistency in a multi-cloud environment • Continuous replication of changing petabyte- scale data without disruption Benefits • Enable heterogenous storage • Access data from anywhere (read and write) • Simplify management of multiple cloud providers • Enable multi-cloud operations • Avoid vendor lock-in • Geo data availability Use Cases • Global data consistency • Guarantee global data accuracy: across clouds, regions, locations, and hybrid deployments • Eliminate reliance on cloud vendors • Data resiliency, downtime, failures, regulations, compliance, performance, cost • Use the right cloud for the right job • Optimize strategy, business performance and operations
  57. 57. © 2019 WANdisco PLC. All rights reserved. Paxos is the foundation for WANdisco’s consensus technology: DConE Correctness Distributed Coordination Continuous Availability Leaderless consensus OptimalityProven DConE: WANdisco’s unique implementation of Paxos has overcome operationalization difficulties. Superior to other consensus approaches particularly in large scale production uses e.g. ZAB (Zookeeper), Raft, Group Communication (Galera) Concurrency and multiple state machines

×