2015 Public Cloud Disaster Recovery Survey

Transcription

1 2015 Public Cloud Disaster Recovery Survey Disaster Recovery Challenges and Best Practices

2 Executive Summary This benchmark survey presents challenges and best practices of companies that host web applications in the public cloud. The results presented here are based on responses from 109 IT professionals from North America and Europe collected through an online survey. Some of the key findings of the survey include: The #1 risk to system availability is Human Error followed by Networks Failures and Cloud Provider Downtime. While the vast majority of the organizations surveyed (83%) have a service availability goal of 99.9% or better, almost half of the companies (44%) had at least one outage in the past 3 months, and over a quarter (27%) had an outage in the past month. The cost of a day of downtime in 37% of the organizations is more than $10,000. When it comes to service availability, there is a clear gap between how organizations perceive their track record and the reality of their capabilities. While almost all respondents claim they meet their availability goals consistently (37%) or most of the time (50%), 28% of the organizations surveyed don t measure service availability at all. It is hard to tell how these organizations claim to meet their goals when they are not able to measure them. The top challenges in meeting availability goals are budget limitations, insufficient IT resources, and lack of in-house expertise. There is a strong correlation between the cost of downtime and the average hours per week invested in backup / disaster recovery. CloudEndure

3 Table of Contents Cloud Infrastructure Provider... 4 Importance of Service Availability... 5 Service Availability Goals... 6 What is downtime?... 7 Meeting Service Availability Goals... 8 Last Downtime Event... 9 Service Availability Measurement Tools and Services Communicating System Availability Statistics to Customers RPO (Recovery Point Objective) and RTO (Recovery Time Objective) Meeting RPO and RTO Goals How Frequent is the System Down for Scheduled Maintenance? Average Monthly Planned Downtime Primary Risks to System Availability Primary Challenges in Meeting Availability Goals Strategies and technologies to Ensure System Availability and Data Protection Frequency of Backup / Replication Application Data Who is Responsible for Disaster Recovery? Average of Hours a Week Invested in Backup / Disaster Recovery Annual Cost of Backup / Disaster Recovery Cost of Downtime Annual Budget for Backup / Disaster Recovery vs. Cost of Downtime Average of Hours a Week Invested in Backup / Disaster Recovery vs. Cost of Downtime Respondent Demographics: Customers Respondent Demographics Respondent Demographics Table of Figures CloudEndure

4 Cloud Infrastructure Provider 59% are hosted in Public Cloud and 74% of the survey respondents use Amazon for their cloud infrastructure. Other cloud infrastructure providers include Microsoft (7%), Google (6%) and Rackspace (4%). 59% Public cloud (e.g. AWS, Azure, Google) 74% 14% 14% 12% Private Cloud Hybrid Cloud Managed Hosting Figure 1: Cloud infrastructure provider 7% 6% 4% 9% Amazon Microsoft Google RackSpace Other Figure 2: Cloud infrastructure provider CloudEndure

5 Importance of Service Availability A vast majority of the survey respondents indicated service availability is highly critical to their customers. 33% rated it most critical, while 71% rated it 8 or higher on a scale of % 27% 11% 13% 7% 2% 2% 2% 2% 0% 10 - Most Critical Not Critical Figure 3: How critical is service availability to your customers? CloudEndure

6 Service Availability Goals The vast majority of the organizations surveyed (83%) have a service availability goal of 99.9% ( three nines or less than 9 hours of downtime a year) or better % or above (less than 5 minutes of downtime a year) 99.99% (less than 53 minutes a year) 99.95% (less than 5 hours a year) 99.9% (less than 9 hours a year) 11% 19% 23% 30% 53% have a goal of 99.99% availability ( four nines or less than 53 minutes a year) or better, and 23% have a five nines availability goal of % or above (less than 5 minutes of downtime a year). 99.8% (less than 18 hours a year) 99.5% (less than 44 hours a year) 99% (less than 88 hours a year) Less than 99% No defined goal 2% 2% 2% 2% 9% Figure 4: Service availability goal CloudEndure

7 What is downtime? 50% of the survey respondents define downtime as inaccessible system. 50% In addition, many of the organizations expand the definition of downtime to include instances when the system is accessible but some functions are not operational (24%) or performance is highly degraded (26%). 26% 24% The system is not accessible The systems is accessible but performance is highly degraded The system is accessible but some functions are not operational Figure 5: Definition of downtime CloudEndure

8 Meeting Service Availability Goals The vast majority of the survey respondents meet their service availability goals consistently (37%) or most of the time (50%). 9% of the respondents indicate that they meet their goals only some of the times. We meet our goals some of the times 9% We are unable to meet our goals 4% We meet our goals consistently 37% We meet our goals most of the time 50% Figure 6: Meeting service availability goals CloudEndure

9 Last Downtime Event While almost all the respondents claim they meet their availability goals consistently or most of the time, almost half of the companies (44%) had an outage in the past 3 months and over a quarter (27%) had an outage in the past month. 6% 21% 17% 13% 15% 19% 9% Week Month 3 months 6 months Year More than a year Never Figure 7: Last downtime event CloudEndure

10 Service Availability Measurement Tools and Services Close to half (49%) of the organizations surveyed use their own tools to measure service availability. 49% 24% use third party tools or services while 28% don t measure service availability at all. 28% 13% 11% We use our own measurement tools We do not measure system availability We use a third party tool We use a third party service Figure 8: Service availability measurement tools and services CloudEndure

11 Communicating System Availability Statistics to Customers Only 15% of the organizations don t share system availability statistics with their customers. 42% Close to a half (42%) notify customers when an event occurs, while 22% share statistics with their customers via regular updates and 20% have a website that shows it. 22% 20% 15% We notify customers We notify customers when an event occurs via regular updates (e.g. monthly) We have a site that shows it We do not share this information with customers Figure 9: Communicating system availability statistics to customers CloudEndure

12 RPO (Recovery Point Objective) and RTO (Recovery Time Objective) 20% of the respondents have a Recovery Point Objective (RPO) of less than 5 seconds data loss. At the other end of the spectrum, 7% have no RPO defined. 11% 9% 7% 17% 26% 24% 7% 0% Zero < 5 seconds 5-30 seconds 1-5 minutes 5-30 minutes minutes More than 1 hour We don't have one Figure 10: RPO Just 9% of the respondents have a Recovery Time Objective (RTO) of less than 5 seconds downtime. 15% have an RTO of more than one hour, and another 7% have no RTO defined. 7% 2% 4% 17% 22% 26% 15% 7% Zero < 5 seconds 5-30 seconds 1-5 minutes 5-30 minutes minutes More than 1 hour We don't have one Figure 11: RTO CloudEndure

13 Meeting RPO and RTO Goals 85% of the organizations meet their Recovery Point Objective consistently (58%) or most of the time (27%). 58% 27% 4% 4% 7% Meet our goal consisently Meet our goals most of the time Meet out goals some of the time Unable to meet our goals N/A Figure 12: Meeting RPO goals 76% meet their Recovery Time Objective consistently (35%) or most of the time (41%). 35% 41% 7% 7% 11% Meet our goal consisently Meet our goals most of the time Meet out goals some of the time Unable to meet our goals N/A Figure 13: Meeting RTO goals CloudEndure

14 How Frequent is the System Down for Scheduled Maintenance? 11% of the survey respondents never have their system down for scheduled maintenance. Never Less than once a month 11% 59% Over half (59%) indicate that their system is down for scheduled maintenance less than once a month. At the same time, 8% of the organizations take their system down for scheduled maintenance every 2 weeks or more frequently. Once every 3-4 weeks Once every 2 weeks Once a week More than once a week 4% 4% 2% 20% Figure 14: How Frequently is the system down for scheduled maintenance? CloudEndure

15 Average Monthly Planned Downtime Two thirds of the organizations surveyed (65%) have less than 30 minutes of monthly planned downtime. 65% 24% have an average of more than one hour monthly planned downtime. 22% 11% 0% 2% Less than 30 minutes minutes 1-5 hours 5-10 hours More than 10 hours Figure 15: Average monthly planned downtime CloudEndure

16 Primary Risks to System Availability According the survey respondents, the top five risks to system availability are: 1. Human Errors 2. Network failures 3. Cloud provider downtime 4. External threats 5. Application scalability limitations Human errors (including Application bugs) Network failures Cloud provider downtime External threats (e.g. hacking, denial of service attack) Application scalability limitations Storage failures Other 1.9 Figure 16: Primary risks to system availability (1-10 scale) CloudEndure

17 Primary Challenges in Meeting Availability Goals The top three primary challenges in meeting availability goals are: 1. Budget limitations 2. Insufficient IT resources 3. Lack of in-house expertise Budget limitations Insufficient IT resources Lack of in-house expertise Limit on our ability to prevent software bugs Coping with peak demand Limit on our ability to prevent configuration errors Cloud resource capacity Keeping up with the pace of change/ growth Figure 17: Primary challenges in meeting availability goals CloudEndure

18 Strategies and technologies to Ensure System Availability and Data Protection Remote storage backup and Storage replication are the leading strategies to ensure system availability and data protection cited by 57% and 46% of the respondents respectively. Remote (multi region/zone) storage backup Storage replication (multi region/zone) Multi-availability region/zone deployment Load balancing 46% 43% 41% 57% Local (single region/zone) storage backup 37% Storage replication (single region/zone) 37% Auto scaling 28% Other 2% Figure 18: Strategies and technologies to ensure system availability and data protection CloudEndure

19 Frequency of Backup / Replication Application Data There are significant variations among organization in the frequency of backup / replication of application data. 34% At one end of the spectrum, 23% apply continuous data replication. 23% 20% At the other end, 34% perform backup every % 7% 5% 2% Continuous Every 5 mins or less Every few minutes (at least Every 1-4 hours Every 4-12 Every hours hours Other Figure 19: Frequency of backup / replication application data CloudEndure

20 Who is Responsible for Disaster Recovery? In over half of the organizations (53%), IT is responsible for disaster recovery. Cloud Operations are the ones in charge of disaster recovery in 27% of the organizations and 13% have a dedicated Business Continuity role. R&D / Application Development 7% Business Continuity 13% Cloud Operations 27% IT 53% Figure 20: Who is responsible for disaster recovery? CloudEndure

21 Average of Hours a Week Invested in Backup / Disaster Recovery 11% of the organizations surveyed don t invest any time at all in backup / disaster recovery. 41% 41% of the organizations invest less than an hour a week in disaster recovery and 35% invest 2-10 hours. 20% 15% 11% 7% 7% None < 1 hour 2-5 hours 5-10 hours > 10 hours Don t know Figure 21: Average of hours a week invested in backup / disaster recovery CloudEndure

22 Annual Cost of Backup / Disaster Recovery 20% of the respondents indicate that their annual cost of backup / disaster recovery is less than $1, % 28% 26% spend $1,000-$10,000 a year, 28% $10,001-$50,000, and 19% spend more than $50,000 a year. 20% 4% 9% 2% 4% 7% < $1K <$10K <$50K <$100K <$500K <$1M $1M+ Don t know Figure 22: Annual cost of backup / disaster recovery CloudEndure

23 Cost of Downtime Only 17% of the organizations surveyed indicated a day of downtime costs $10,000 or less. 30% The cost of a day of downtime in 37% of the organizations is more than $10, % 11% 13% 15% 7% 2% 4% < $1K <$10K <$50K <$100K <$500K <$1M $1M+ Don t know Figure 23: Cost of one day of downtime CloudEndure

24 Annual Budget for Backup / Disaster Recovery vs. Cost of Downtime Disaster Recovery budget is highly correlated to the cost of downtime. 100% 90% 80% 10% 24% 29% 91% of the organizations where the cost of a day of downtime is less than $10,000 have an annual backup / disaster recovery budget of less than $10,000, whereas 70% of the organizations with a day of downtime cost exceeding $10,000 have an annual backup / disaster recovery budget of over $10, % 60% 50% 40% 30% 20% 10% 0% 41% 67% 29% < $10,000 > $10,000 Cost of one day of downtime Annual Budget for backup/ DR > $100,000 $10,000 - $100,000 < 10,000$ Figure 24: Annual budget for backup / disaster recovery vs. cost of one day of downtime CloudEndure

25 Average of Hours a Week Invested in Backup / Disaster Recovery vs. Cost of Downtime There is a strong correlation between the cost of downtime and the average hours a week invested in backup / disaster recovery. 68% of the organizations where the cost of a day of downtime is less than $10,000 invest less than an hour in average in backup / disaster recovery, while 35% of the organizations with a daily downtime cost exceeding $100,000 invest over 5 hours in backup / disaster recovery. 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 8% 20% 24% 35% 42% 47% 68% 38% 18% < 10,000$ $10,000 - $100,000 > $100,000 Cost of one day of downtime Annual Budget for backup/ DR > 5 hours 2-5 hours < 1 hour or none Figure 25: Average hours a week invested in backup / disaster recovery vs. cost of one day of downtime CloudEndure

26 Respondent Demographics: Customers The majority of survey respondents serve more than one type of customers. 46% of survey respondents serve consumers and 48% serve small businesses, while 46% serve enterprises. Consumers 46% SMB 48% Enterprise 54% Figure 26: Type of customers served CloudEndure

27 Respondent Demographics 43% of the survey respondents come from organizations of up to 50 employees, 42% have 51-1,000 employees and 16% over 1, % 35% 7% 7% 9% , > 5,000 Figure 27: Number of employees 52% of the respondents have less than $10 million in revenues, 26% have $10-100M and 22% over $100M. 52% 13% 13% 15% 7% <$10M $10-50M $51-100M $101M-$1B >$1B Figure 28: Annual revenues CloudEndure

28 Respondent Demographics 20% of the survey respondents are responsible for application development and another 20% each for Cloud and IT operations. Compliance 2% Disaster Recovery 2% Database 2% Development Opera'ons 7% Infrastructure (cross- domain) 16% Other 11% Applica'on Development 20% IT Opera'ons 20% Cloud 20% Figure 29: Job responsibility 46% of the respondents are in Director or Manager positions and 19% are VPs or C- Level. Team Member 12% Vice President 5% Supervisor 2% Director 23% C- level 12% Team Leader 23% Manager 23% Figure 30: Position in the organization CloudEndure

29 Table of Figures Figure 1: Cloud infrastructure provider... 4 Figure 2: Cloud infrastructure provider... 4 Figure 3: How critical is service availability to your customers?... 5 Figure 4: Service availability goal... 6 Figure 5: Definition of downtime... 7 Figure 6: Meeting service availability goals... 8 Figure 7: Last downtime event... 9 Figure 8: Service availability measurement tools and services Figure 9: Communicating system availability statistics to customers Figure 10: RPO Figure 11: RTO Figure 12: Meeting RPO goals Figure 13: Meeting RTO goals Figure 14: How Frequently is the system down for scheduled maintenance? Figure 15: Average monthly planned downtime Figure 16: Primary risks to system availability (1-10 scale) Figure 17: Primary challenges in meeting availability goals Figure 18: Strategies and technologies to ensure system availability and data protection Figure 19: Frequency of backup / replication application data Figure 20: Who is responsible for disaster recovery? Figure 21: Average of hours a week invested in backup / disaster recovery Figure 22: Annual cost of backup / disaster recovery Figure 23: Cost of one day of downtime Figure 24: Annual budget for backup / disaster recovery vs. cost of one day of downtime Figure 25: Average hours a week invested in backup / disaster recovery vs. cost of one day of downtime Figure 26: Type of customers served Figure 27: Number of employees Figure 28: Annual revenues Figure 29: Job responsibility Figure 30: Position in the organization CloudEndure

30 About CloudEndure Business as Usual. Always. CloudEndure provides Cloud Migration and Cloud Disaster Recovery for any application, allowing companies to mobilize entire applications with their data to and across clouds with near zero downtime and no data loss. CloudEndure enables truly consistent, block-level, real-time replication using continuous data protection (CDP). Founded in 2012, CloudEndure s Cloud Workload Mobility technology creates an exact copy of the entire application at an alternative cloud location at the touch of a button, within minutes, and with the latest data. CloudEndure supports physical, virtualized or cloud-based applications as the source and Amazon Web Services (AWS), Google Cloud Platform (GCP) and Microsoft s Azure as target cloud locations. For more information, visit Survey sponsors: CloudEndure