F5, Inc. · NASDAQ:FFIV

F5 — F5 Networks — Disaster Recovery: Data Replication and the WAN webinar (March 2007)

56 slides from F5's conference presentation, Mar 21, 2007

  1. Slide 1: Disaster Recovery: Data Replication and the WAN

    1 Disaster Recovery: Data Replication and the WAN Presented by: Stephanie Balaouras, Senior Analyst, Forrester Research Kevin Hohenbrink, Product Manager, F5 Networks

  2. Slide 2: The Impact Of The WAN On DR Capabilities

    The Impact Of The WAN On DR Capabilities Stephanie Balaouras Senior Analyst Forrester Research March 21st, 2007

  3. Slide 3: Agenda

    Agenda • Survey fielding and methodology • Current state of disaster recovery preparedness • Enterprise demand for improved recovery capabilities • Impact of the WAN on recovery capabilities • Market awareness of WAN optimization technologies to facilitate disaster recovery 3 Entire contents © 2007

  4. Slide 4: Survey fielding and methodology

    Survey fielding and methodology • 504 web-based surveys • Respondents » Decision-makers and influencers in planning for and/or purchasing BC/DR related solutions – CIO, VPs, Directors, and Managers – IT Operations, Enterprise Architecture, Network • Geographic focus: » 200 North America: US, Canada

  5. Slide 5: Enterprises spend 21% of the IT budget on BC/DR

    Enterprises spend 21% of the IT budget on BC/DR “Using your best estimate, what % of your IT budget is spent on BC/DR?” 35 30 25 North America Europe Percent 20 15 10 5 0 less 10% - 20% - 30% - 40% - 50% - 60% - 70% - than <20% <30% <40% <50% <60% <70% <80% 10% Base: North America 200, Europe 304 So

  6. Slide 6: 2/3 of enterprises have more than one DR site

    2/3 of enterprises have more than one DR site “Do you have a backup data center?” 63% Yes, two or more 67% Europe North America 37% Yes, one 34% Base: North America 200, Europe 304 Source: Disaster recovery and data replication study conducted by Forrester Consulting and commissioned by F5 Networks,

  7. Slide 7: Enterprises with multiple DR sites prefer multi-hop or bi-directional DR configurations

    Enterprises with multiple DR sites prefer multi-hop or bi- directional DR configurations “What is your DR configuration?” 24% Concurrent 25% 40% Europe Multi-hop 35% North America Two site bi- 36% directional 41% Base: North America 133, Europe 193 Respondents with more than one back-up data center

  8. Slide 8: Enterprises favor much shorter distances between sites

    Enterprises favor much shorter distances between sites “What is the distance between your primary and backup data center (s)?” Europe North America > 8000 km 4% 14% 6000 - < 8000 km 2% 5% 4000 - < 6000 km 3% 15% 2000 - < 4000 km 6% 20% 500 - < 2000 km 27% 22% < 500 km 58% 24% Base: North America 200

  9. Slide 9: Europe uses 500 Mbps to 2.5 Gbps between data centers

    Europe uses 500 Mbps to 2.5 Gbps between data centers On average, how much wide-area network bandwidth do you have between your primary data center and your backup data center(s)? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% To Backup Data Center 7 14 13 28 17 12 10 Between Bi-directional Sites 3 4 2

  10. Slide 10: North America relies on much more bandwidth between sites

    North America relies on much more bandwidth between sites On average, how much wide-area network bandwidth do you have between your primary data center and your backup data center(s)? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% To Backup Data Center 3 10 28 21 19 13 5 Between Bi-directional Sites 0

  11. Slide 11: Ethernet is the dominant transport in North America

    Ethernet is the dominant transport in North America What type of wide-area network transport do you use between the primary data center and your backup data center(s)? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% To Backup Data Center 18 12 40 5 13 6 2 4.5 Between Bi-directional Sites 15 20 24 15 2 1

  12. Slide 12: Internet and Ethernet dominant transport in Europe

    Internet and Ethernet dominant transport in Europe What type of wide-area network transport do you use between the primary data center and your backup data center(s)? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% To Backup Data Center 26 5 48 5 10 23 Between Bi-directional Sites 26 9 45 9 3 3 6 Multi

  13. Slide 13: Most enterprises can recover operations in 2-5 hrs

    Most enterprises can recover operations in 2 - 5 hrs “In the event of a primary data center site failure, “If your expected recovery time is less than 2 hours, what is your expected recovery time in hours?” please estimate your recovery time in minutes?” Europe North America > 30 hours 8% 0% 8% 100

  14. Slide 14: Most enterprises will lose between 2-5 hrs of data

    Most enterprises will lose between 2 – 5 hrs of data “In the event of a primary data center site failure, “If your expected data loss is less than two hours, please how many hours of data will you lose?” estimate your recovery time in minutes?” Europe North America > 30 hours 6% 4% 5% 100 - < 120 mi

  15. Slide 15: The amount of data to be replicated is increasing

    The amount of data to be replicated is increasing “Please estimate the amount of data (in terabytes) that must be protected at the primary site?” 24% > 500 TB 26% 2% 400 - < 500 TB 3% 2% 300 - < 400 TB 6% 3% 200 - < 300 TB 5% Europe 100 - < 200 TB 13% North America 9% 9% 50 - < 100 TB 9% 10% 25 - <

  16. Slide 16: Bandwidth costs 29% of remote replication or backup solutions

    Bandwidth costs 29% of remote replication or backup solutions “What percent of the total cost of your replication or remote backup solution can you attribute to the cost of bandwidth?” 70% - <80% 0% 1% Europe 60% - <70% 4% North America 3% 50% - <60% 4% 3% 40% - <50% 9% 5% 30% - < 40% 13% 8% 20% - <

  17. Slide 17: Majority believe it is critical to improve RTO and RPO

    Majority of enterprises believe its critical or very critical to improve RTO and RPO “On a scale of 1-4, where 1 is not critical and 4 is very critical, how important is it to improve your time to recovery at the backup data center and to limit data loss?” Not Critical Somewhat Critical Critical Ver

  18. Slide 18: Cost of downtime and increased risk drive improved recovery in North America

    Cost of downtime and increased risk drive improved recovery in North America “What is driving the need to improve time to recovery and to limit data loss? Rank the top 3?” Cost of downtime 73 44 28 Fiduciary responsibility 45 35 33 Increased risk 42 46 46 Requirement to stay online and competitive 2

  19. Slide 19: Cost of downtime and competition drive recovery improvements in Europe

    Cost of downtime and competition drive recovery improvements in Europe “What is driving the need to improve time to recovery and to limit data loss? Rank the top 3?” Cost of downtime 101 85 41 Fiduciary responsibility 63 35 46 Increased risk 55 33 33 Requirement to stay online and competitive 24X7 5

  20. Slide 20: Unprotected sites represent a significant risk

    Unprotected sites represent a significant risk What % of remote sites replicate or backup data to data center or central facility? 15% 80% to 100% 16% 13% 60% to 80% 20% 20% Europe 40% to 60% 24% North America 31% 20% to 40% 22% 21% less than 20% 18% Base: North America 200, Europe 304 Source: Disas

  21. Slide 21: Impact of bandwidth on North American recovery objectives

    Impact of bandwidth on North American recovery objectives On a scale of 1 to 4, where 1 equals “no impact” and 4 equals “very strong impact”, please rate the impact of the following to your ability to improve your time to recovery and to limit data loss. No Impact Somewhat Impactful Impactful Very I

  22. Slide 22: Impact of bandwidth on European recovery objectives

    Impact of bandwidth on European recovery objectives On a scale of 1 to 4, where 1 equals “no impact” and 4 equals “very strong impact”, please rate the impact of the following to your ability to improve your time to recovery and to limit data loss. No Impact Somewhat Impactful Impactful Very Impactf

  23. Slide 23: 86% of North American enterprises want to improve recovery without increasing bandwidth

    86% of North American enterprises want to improve recovery without increasing bandwidth Please rate your level of agreement with the following statements regarding your WAN bandwidth. (1 equals strongly disagree and 4 equals strongly agree) Don't Agree Somewhat Agree Agree Strongly Agree Our current

  24. Slide 24: 79% of European enterprises want to improve recovery without increasing bandwidth

    79% of European enterprises want to improve recovery without increasing bandwidth Please rate your level of agreement with the following statements regarding your WAN bandwidth. 1 equals strongly disagree and 4 equals strongly agree. Don't Agree Somewhat Agree Agree Strongly Agree Our current bandwi

  25. Slide 25: Awareness of WAN optimization solutions is high

    Awareness of WAN optimization solutions is high Please rate your level of awareness of WAN acceleration appliances as a way to improve the performance of replication and remote backup. (On a scale of 1-4, where 1 is not at all aware and 4 is very aware) Not Aware Somewhat Aware Aware Very Aware Euro

  26. Slide 26: Recommendations

    Recommendations • Don’t selectively replicate only the most mission-critical applications. » Business processes rely on multiple integrated applications • Protect your remote sites • Before investing in additional bandwidth to support remote sites, consider WAN optimization offerings • When evaluati

  27. Slide 27: Thank you

    Thank you Stephanie Balaouras +1 617/613-6440 [email protected] www.forrester.com 27 Entire contents © 2007 Forrester Research, Inc. All rights reserved.

  28. Slide 28: Data Replication: F5 and WAN Optimization

    28 Data Replication: F5 and WAN Optimization

  29. Slide 29: Who is F5?

    29 Who is F5? 100 90 Service Product 80 Public company FFIV (Nasdaq) 70 Founded 1996 60 FY’06 revenue $394.0 million 50 40% YoY growth $ Millions 40 1,000+ employees 30 Products: 20 – Application Delivery Networking 10 – Network Optimization – Network and Application Security 0 3Q03 4Q03 1Q04 2Q04 3

  30. Slide 30: Corporate Customers

    30 Corporate Customers 9 of top-10 Global Finance firms, and 60% of the Global 1000 Financial IT Telco/ISP Media Transport Other

  31. Slide 31: Common Symptoms of Replication with Network Problems

    31 Common Symptoms of Replication with Network Problems Unable to meet storage SLA’s (RPOs and RTOs) – Expanding distance between DR sites Performance and uptime problems Expectation of need to purchase more bandwidth Root Causes: – Bandwidth is limited – Latency, packet-loss, congestion variable –

  32. Slide 32: How Can The Problem Be Solved?

    32 How Can The Problem Be Solved? 1. Replicate less data? Primary Site Backup Site WAN

  33. Slide 33: How Can The Problem Be Solved?

    33 How Can The Problem Be Solved? 1. Replicate less data? NO! Primary Site Backup Site WAN

  34. Slide 34: How Can The Problem Be Solved?

    34 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Primary Site Backup Site WAN

  35. Slide 35: How Can The Problem Be Solved?

    35 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Expensive, recurring, ineffective Primary Site Backup Site WAN

  36. Slide 36: How Can The Problem Be Solved?

    36 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Expensive, recurring, ineffective 3. Accelerate DR traffic? Primary Site Backup Site WAN

  37. Slide 37: How Can The Problem Be Solved?

    37 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Expensive, recurring, ineffective 3. Accelerate DR traffic? Cost effective solution Primary Site Backup Site WAN

  38. Slide 38: How Can The Problem Be Solved?

    38 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Expensive, recurring, ineffective 3. Accelerate DR traffic? Cost effective solution 4. Prioritize DR traffic and guarantee bandwidth? Primary Site Backup Site WAN

  39. Slide 39: How Can The Problem Be Solved?

    39 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Expensive, recurring, ineffective 3. Accelerate DR traffic? Cost effective solution 4. Prioritize DR traffic and guarantee bandwidth? Most effective when combined with #3 Primary Site Backup Site WAN

  40. Slide 40: How Can The Problem Be Solved?

    40 How Can The Problem Be Solved? 1. Replicate less data? NO! 2. Lease more WAN bandwidth? Expensive, recurring, ineffective 3. Accelerate DR traffic? Cost effective solution 4. Prioritize DR traffic and guarantee bandwidth? Most effective when combined with #3 Primary Site Backup Site WAN

  41. Slide 41: Solution: F5 WANJet WAN Optimization

    41 Solution: F5 WANJet WAN Optimization 2 years in the market Rack-mountable appliance Enables LAN-like performance across a WAN Symmetric deployment architecture (one on each side of the WAN)

  42. Slide 42: WANJet Value Proposition

    42 WANJet Value Proposition Meet RPOs and RTOs without upgrading bandwidth or Data Replication (DR) infrastructure Accelerate DR replication up to 5 to 10 times faster* Utilize 70-90% less bandwidth* Guarantee bandwidth and prioritize DR traffic over non-DR traffic Mitigate the effects of latency En

  43. Slide 43: WANJet Value Proposition

    43 WANJet Value Proposition Meet RPOs and RTOs without upgrading bandwidth or Data Replication (DR) infrastructure Accelerate DR replication up to 5 to 10 times faster* Utilize 70-90% less bandwidth* Guarantee bandwidth and prioritize DR traffic over non-DR traffic Mitigate the effects of latency En

  44. Slide 44: WANJet Value Proposition

    44 WANJet Value Proposition Meet RPOs and RTOs without upgrading bandwidth or Data Replication (DR) infrastructure Accelerate DR replication up to 5 to 10 times faster* Utilize 70-90% less bandwidth* Guarantee bandwidth and prioritize DR traffic over non-DR traffic Mitigate the effects of latency En

  45. Slide 45: WANJet Value Proposition

    45 WANJet Value Proposition Meet RPOs and RTOs without upgrading bandwidth or Data Replication (DR) infrastructure Accelerate DR replication up to 5 to 10 times faster* Utilize 70-90% less bandwidth* Guarantee bandwidth and prioritize DR traffic over non-DR traffic Mitigate the effects of latency En

  46. Slide 46: WANJet Value Proposition

    46 WANJet Value Proposition Meet RPOs and RTOs without upgrading bandwidth or Data Replication (DR) infrastructure Accelerate DR replication up to 5 to 10 times faster* Utilize 70-90% less bandwidth* Guarantee bandwidth and prioritize DR traffic over non-DR traffic Mitigate the effects of latency En

  47. Slide 47: How Latency Effects are Mitigated by WANJet

    47 How Latency Effects are Mitigated by WANJet Transparent Data Reduction #2 (byte-level pattern matching) Reduce the amount of data transmitted Transparent Data Reduction #1 (compression) Latency has less direct Transmit more efficiently impact on DR performance TCP Acceleration Manage limited band

  48. Slide 48: How Latency Effects are Mitigated by WANJet

    48 How Latency Effects are Mitigated by WANJet Transparent Data Reduction #2 (byte-level pattern matching) Reduce the amount of data transmitted Transparent Data Reduction #1 (compression) Latency has less direct Transmit more efficiently impact on DR performance TCP Acceleration Manage limited band

  49. Slide 49: What Is Total Effective Latency

    49 What Is Total Effective Latency Total Network Other Latency- Effective = Latency + Inducing Factors Latency This is what impacts Data Replication performance This is determined by point-to-point distance, and the speed of light. This artificial latency is caused by many factors including: • Conge

  50. Slide 50: WanJet Data Replication Prioritization & Bandwidth Allocation

    50 WanJet Data Replication Prioritization & Bandwidth Allocation (a) Prioritize DR traffic over non-DR traffic (by port or IP address) (b) Allocate minimum bandwidth to Data Replication to protect it from high-congestion situations Mbps Throughput Coarse, nailed up pipe capacity Bandwidth Savings ($

  51. Slide 51: WANJet Changes to Data Replication traffic: None

    51 WANJet Changes to Data Replication traffic: None Transparent to Data Replication Traffic WANJet acts as an inline accelerator. It does not alter or redirect Data Replication traffic. WANJet “fails-to-wire” (i.e. traffic keeps passing through) Multiple WANJets can be load balanced for high availab

  52. Slide 52: Sample Configuration

    52 Sample Configuration Example Below: 2.4 Gbps data throughput over a 300 Mbps WAN with 8X Compression and redundant failover architecture N+1 Scalability with Failover Capacity F5 WANJets F5 WANJets Storage Storage on SRDF on SRDF BIG-IP Load Balancing with Failover Capacity F5 BIG-IP’s F5 BIG-IPs

  53. Slide 53: Case Study: US Government Agency

    53 Case Study: US Government Agency WANJet CHALLENGES SRDF/A on EMC storage 1,000+ miles between sites 18 Mb/sec of bandwidth provided, but over 90 Mb/sec needed based on 2:1 compression rate to run replication RESULTS 5:1 Optimization/Compression rate achieved SRDF/A replication able to run on 18 M

  54. Slide 54: Case Study: Fortune 10 Financial

    54 Case Study: Fortune 10 Financial WANJet CHALLENGES Replication Performance 20 mile, Dual T3 Links Weekly copy back-up took 9 Days RESULTS Weekly copy reduced to 2 Days 2/3 of bandwidth usage saved Application Response Near Real-Time Payback period in less than 6 mo’s Avoided potential bandwidth u

  55. Slide 55: For More Information

    55 For More Information www.f5.com – Independent report: Forrester Impact of the WAN on Disaster Recovery Capabilities – Solution Center: Optimization for white papers, success stories – WANJet product section – WANJet demo-on-demand – Podcasts: www.f5.com/communication/podcasts.html www.f5resources

  56. Slide 56: The World Runs Better With F5

    5 THE WORLD RUNS BETTER WITH F5 56

F5 Networks — Disaster Recovery: Data Replication and the WAN webinar (March 2007) on 6ix