Cloud vs. On-Premise Disaster Recovery: A Comparison

Part One of a Multi-Part Series on Disaster Recovery in Cloud

Table of Contents 

• Basics of Disaster Recovery Planning
• On-Premise vs. Cloud Disaster Recovery
• Importance of Planning for Cloud Disasters
• Conclusion
• Jira Disaster Recovery in Action - Walkthrough Video
• Useful Links and References  

Disaster recovery planning is essential for any organization to ensure business continuity and minimize data loss in the event of an unforeseen disaster. Traditionally, these plans were centered around on-premise infrastructure. With the advent of cloud computing, the landscape of disaster recovery has transformed, bringing both new opportunities and challenges.

Basics of Disaster Recovery Planning

A disaster recovery (DR) plan outlines the processes, policies, and procedures necessary to recover and protect an organization’s IT infrastructure in the event of a disaster.

Key elements include:

• Risk Assessment and Business Impact Analysis: Identifying potential risks, such as natural disasters, cyber-attacks, or hardware failures, and understanding their impact on business operations.
• Recovery Time Objectives (RTO): The maximum acceptable length of time that your application can be offline. This defines how quickly systems must be restored after a disaster to minimize downtime and financial losses.
• Recovery Point Objectives (RPO): The maximum acceptable amount of data loss measured in time. This determines the frequency of data backups required to ensure that data loss is minimized.
• Data Backup and Restoration: Regularly backing up data and ensuring that backups are stored securely and can be restored quickly in the event of data loss.
• DR Site and Infrastructure: Establishing a secondary site (either on-premise or in the cloud) that can take over operations if the primary site fails.
• Communication Plan: Defining clear communication channels and procedures for informing stakeholders, employees, and customers about the disaster and recovery process.
• Testing and Maintenance: Regularly testing the DR plan to ensure it works as expected and updating it to address new risks or changes in the business environment.
  
  

Do I Even Need A Plan In Cloud?

There are two key reasons why everyone who is using cloud applications for mission critical business operations needs a disaster recovery plan.  They are;

1) Disasters Happen to EVERYONE.  If you are not convinced about this, then scroll to the bottom of this article and check out the references and read this article  How Google Deleted $125Billion Dollar Fund 

2) Most large cloud providers have a Shared Responsibility Model which pushes part of the responsibility of data resiliency back on the end customer, that's you. 

On-Premise vs. Cloud Disaster Recovery

The following is a fairly generic overview of on-premise vs. cloud disaster recovery topics.  In future articles in this series, we will dive deep into the nuances of how these topics apply to the Atlassian Cloud ecosystem and we will assess a variety of tools and approaches in order to meet your disaster recovery objectives.

Risk Assessment and Business Impact Analysis

• On-Premise: Requires identifying risks specific to physical locations, such as natural disasters and power outages.
• Cloud: Focuses on cloud-specific risks like service outages, data breaches, and vendor lock-in. For instance, cloud environments are not immune to outages as evidenced by past incidents with AWS and Azure.

Recovery Time Objectives (RTO)

• On-Premise: Achieving low RTOs involves significant investment in redundant hardware and swift manual recovery processes.
• Cloud: Cloud providers offer automated recovery solutions that can drastically reduce RTOs. Services like AWS Elastic Disaster Recovery can automate the failover process, ensuring rapid recovery.
  One critical factor in RTO for SaaS applications is how fast the vendor is able to recover the platform, or the feasibility of the customer restoring data and systems to another environment.  

Recovery Point Objectives (RPO)

• On-Premise: Regular physical backups are essential, often requiring manual intervention and physical storage solutions.
• Cloud: Cloud services offer continuous data replication across geographically diverse data centers, significantly reducing RPO. For example, using services like AWS S3 or Azure Blob Storage, businesses can ensure near-real-time data replication.
  
  The world of SaaS on the other hand offers a lot of challenges with backups as backup software vendors are usually limited to the data access that they have by the vendor's API rules.  Thus, flexibility in RPO can be challenging. 

Data Backup and Restoration

• On-Premise: Involves managing physical backup media and ensuring secure off-site storage, which can be logistically challenging and costly.
• Cloud: Cloud storage solutions provide automated, scalable, and secure backup options. Cloud backup services eliminate the need for physical media, reducing the risk of data loss and simplifying restoration.

DR Site and Infrastructure

• On-Premise: Setting up and maintaining a secondary physical site can be expensive and complex.
• Cloud: Cloud infrastructure enables the use of multi-region deployments and elastic resources, allowing businesses to quickly scale up their DR site without the need for significant capital expenditure.

Communication Plan

• On-Premise: Relies heavily on internal communication systems, which may be disrupted during a disaster.
• Cloud: Leverages cloud-based communication tools that can remain operational even if the primary site is compromised. Services like Amazon SNS or Azure Notification Hubs can automate notifications to stakeholders.

"Have a Plan, But Practice It"
Sarah Wright during "Jira is Down" livestream of The Jira Life

Testing and Maintenance

• On-Premise: Regular testing can be resource-intensive and may require taking systems offline.
• Cloud: Cloud providers offer testing environments and simulation tools to regularly test and update DR plans without disrupting live systems. For instance, AWS Fault Injection Simulator can help in testing the resilience of cloud environments.

Importance of Planning for Cloud Disasters

Despite the many advantages of cloud-based disaster recovery, it is crucial to understand that moving to the cloud does not eliminate the need for a robust DR plan.

Here are specific examples highlighting why cloud DR planning is essential:

1. Security Risks: “Our tests succeeded because the Department failed to implement security measures capable of either preventing or detecting well-known and widely used techniques employed by malicious actors to steal sensitive data” (TechCrunch, source). This example underscores the importance of ensuring that cloud environments are properly secured against cyber threats.
   
   
2. Insider Threats: “Upset that he was fired, a former employee hacked into his company’s computer system and deleted 20 virtual servers, causing the company to lose S$918,000” (Channel News Asia, source). This incident highlights the risk of internal threats and the necessity of having stringent access controls and monitoring.
   
   
3. Service Outages: "According to InformationWeek’s 2024 Cloud Computing Report, 60% of IT decision-makers use Amazon Web Services (AWS) and half of these respondents have faced disruption due to an outage at AWS in the past 12 months" (DataCenterKnowledge, source). Service disruptions can have significant impacts on business operations, making it critical to have DR plans that account for potential cloud service outages.
   
   
4. Technical Glitches: "Early in 2023, Microsoft experienced a three-hour outage of its core M365 offerings due to Azure network issues, wiping out some of its most popular services" (DataCenterKnowledge, source). Technical issues in cloud infrastructure can lead to downtime, emphasizing the need for comprehensive DR planning.

Conclusion

Disaster recovery planning is a critical aspect of ensuring business continuity and minimizing data loss in the face of unforeseen events. While cloud infrastructure offers significant advantages over traditional on-premise solutions, such as cost efficiency, scalability, and ease of access, it also introduces unique challenges and risks. Effective cloud disaster recovery requires robust security measures, regular testing, and continuous updates to address evolving threats. As we continue this series, we will explore these topics in greater depth, providing actionable insights and best practices to help you safeguard your cloud-based systems. Stay tuned for the next part of our series, where we will delve into identifying and mitigating cloud-specific risks.

Jira Disaster Recovery in Action

We asked Alex Ortiz to help create a step by step disaster recovery video in Jira using our award winning Revyz Data Manager for Jira product.  Watch it below.

Useful Links and References

• A government watchdog hacked a US federal agency to stress-test its cloud security
• Fired employee accessed company’s computer 'test system' and deleted servers, causing it to lose S$918,000
• A History of AWS Cloud and Data Center Outages
• A History of Microsoft Azure Outages