How to restore your critical applications, data, and infrastructure from any source regardless of machine type, disk type, hypervisor, platform, or geography to Oracle Cloud using ZConverter’s price disruptive solution.
Achieve resilience on Oracle Cloud with ZConverter’s all-in-one offering for cloud backup, disaster recovery, migration, and ransomware protection
Planning for Disaster Recovery (“DR”) is like buying insurance: people want to pay the lowest annual premium and when disaster strikes, they panic and wish that they had bought the right type and amount of coverage.
Be it a small-to-medium sized business seeking financial risk reduction, or a Fortune 500 global corporation focused on reliability, they all need a sound business continuity (“BC”) and resiliency strategy so that if disaster strikes, they won’t incur serious negative impacts; financial or otherwise. Many organizations are looking for a reliable and cost-effective DR solution that takes advantage of the on-demand provisioning and pay-as-you-go billing capabilities of the Cloud.
This article discusses the Cloud Disaster Recovery use case and how to utilize ZConverter’s solution to successfully backup and recover workloads to Oracle Cloud.
What is cloud disaster recovery?
Cloud Disaster Recovery (“CDR”) protects valuable business resources and ensures that the organization’s critical systems, applications and data can be accessed and recovered into a target location after a disaster, to resume normal business operations. A viable Cloud DR solution must offer next-generation capabilities that make DR accessible to organizations that previously could not afford it. To deliver this, CDR must be a true “server-less” cloud-based solution i.e., it must not require pre-provisioned servers to be maintained in the target environment to recover workloads and IT services. Next-generation CDR has become a key element of any business continuity strategy due to the flexibility and disruptive pricing it offers.
Figure 1 illustrates a server-less hybrid-cloud and multi-cloud next-generation CDR deployment.
Comparing cloud disaster recovery to traditional solutions
We must distinguish between the following solutions:
Cloud Disaster Recovery (“CDR”)
High Availability (“HA”)
Traditional On-premises Disaster Recovery (“DR”)
Disaster Recovery as-a-Service (“DRaaS”)
A common IT myth is that traditional HA and DR solutions serve the same purpose as CDR solutions. Untrue, they are not the same. Traditional HA solutions can become expensive when they include complex data mapping and replication technology. HA solutions a.k.a. HADR, deliver zero-data-loss-zero-downtime protection for databases. They do not provide ransomware protection, which is achieved via comprehensive backup with point-in-time Recovery Point Objective (“RPO”) and immutable storage. Traditional HA, being very different from DR and CDR, and being expensive, lends itself well to deployment in conjunction with a low-cost CDR solution. In this scenario, the add-on CDR technology provides an extra layer of protection for databases that require zero-data-loss-zero-downtime HA and need ransomware protection with long-term incremental rollback.
Traditional DR is an inflexible and expensive on-premises solution with the high cost of duplicating IT infrastructure in a source location and a fixed target datacenter location. Also, it cannot function over the WAN or migrate servers between disparate environments, and so it requires a dedicated circuit between the source and target datacenters, to operate like a single interconnected LAN environment. Traditional DR also cannot migrate servers between disparate heterogenous environments, such as an on-premises environment and a cloud platform of your choice, or between two different cloud platforms.
DRaaS is usually a service offered by a professional services organization; with a patchwork of vendor-supplied backup solutions along with open-source migration tools to create a tightly controlled and highly specific environment, where the end-user can only recover workloads into the DRaaS provider’s traditional hosting environment from their VMware on-premises environment. These solutions can be expensive and limited in the range of workloads they can protect and the compute environments they can support.
The case for cloud disaster recovery
Consider the example of organizations that have experienced outages due to natural disasters, IT-operator or service-provider errors, data corruption and/or ransomware attacks, or have literally been de-platformed. As a result of these incidents, many have suffered catastrophic losses, some were replaced by opportunistic competitors, and others lost their CEOs to infamy. Why did they have such disastrous outcomes? They either failed to recover in a timely fashion, or they lost large amounts of critical transactional and customer information, or both. They lost data, they lost trust, and they lost customers.
The root cause was that these organizations had not implemented a reliable DR solution. Many organizations find it cost-prohibitive to implement a traditional DR solution that requires procurement and maintenance of an exact replica of the infrastructure in their source and target locations, connected by dedicated leased circuits. So, the million-dollar question is this: is an expensive, zero-data-loss-zero-downtime DR solution absolutely needed to achieve adequate resiliency? Often, the answer is “no”.
The good news is that with a reliable and truly “server-less” CDR solution most organizations can now afford protection that will significantly limit their losses when a disaster occurs. The low recovery-error-rate, reasonable Recovery Time Objective (“RTO”) and low total cost of ownership (“TCO”) of a next-generation CDR is clear. Therefore, as word gets out, CDR will gain widespread adoption by organizations of all sizes to protect their workloads.
In addition to offering cloud backup, CDR addresses the difficult challenge of recovering production workloads: applications, data, and associated IT services such as operating systems, drivers, dependencies, etc., into the cloud, simultaneously. The significantly low cost of next-generation CDR makes DR with ransomware protection more affordable to organizations. CDR is a game changer because it delivers inexpensive advanced technology that empowers organizations with a reliable business continuity strategy.
What makes a good cloud disaster recovery solution?
The main indicators of a good CDR solution are that it is truly “server-less” and delivers highly reliable error-free recovery capability. There are varying degrees of “server-less” and the best CDR solutions eliminate the cost of pre-provisioning servers in the target environment and reduce the overall infrastructure costs such as by eliminating common infrastructure requirements of legacy CDR solutions including: replication servers, conversion servers, dedicated persistent block devices for each source disk, load balancers, bastion hosts, etc.
Further cost reduction is achieved if the CDR replication process requires replication of only the used portion of each source disk and not the entire provisioned disk and is able to store the full and incremental backups in secure low-cost immutable object-based storage. Being truly “server-less” significantly lowers the overall TCO of a CDR solution and is its main advantage over traditional DR due to the on-demand provisioning and pay-as-you-go billing capabilities of the Cloud.
A next-generation CDR solution should deliver the following key technical capabilities:
On-premises and cloud backup, cloud migration and cloud disaster recovery, under one CDR software subscription.
Error-free, any-to-any, source-agnostic migration, and recovery of applications, data, operating systems, associated IT services, and necessary dependencies, regardless of the source and target computing platform, machine type, hypervisor, disk type, disk sizing, disk configuration, disk format, disk partitioning, BIOS/UEFI mode, 32/64-bit system, Intel/AMD CPU, file-format, or datastore type, simultaneously, into any target cloud environment.
“Server-less” capability.
Ransomware protection.
Support for newest and legacy Operating System versions.
Compression and encryption of operating systems, applications, and data.
Low I/O overhead replication-processing.
Replication of only the used portion of source disk.
No requirement for new full backups whenever source servers undergo hardware-level configuration changes or OS patches or upgrades.
Scalable replication-processing options for handling large amounts of data or in write-intensive environments.
Automated server provisioning in target environment to speed up the recovery process.
On-Demand server migration; with no need to wait for a disaster to strike.
Be aware that without these capabilities a CDR solution can be rendered virtually useless due to excessive I/O overhead negatively affecting source-server application performance, high recovery-error-rates, runaway storage and WAN costs, vulnerability to ransomware attacks and burdensome IT management overhead.
Figure 2 illustrates on-premises and cloud backup, cloud migration, and on-premises-to-cloud DR and cloud-to-cloud DR deployments.
What cloud disaster recovery doesn’t offer
While CDR presents a compelling new possibility for ensuring business continuity, it is not everything to everybody. When comparing technology solutions there are always trade-offs. For instance, the RTO for CDR can be longer than with traditional HA and DR, due to CDR’s dependence upon the quality of the WAN connectivity at the source and target locations. Generally, the achievable RTO for traditional HA and DR varies from instantaneous to less than an hour, whereas CDR RTO ranges from a few minutes for a few VMs to possibly a few days for thousands of VMs. CDR RTO depends mostly on the backup size being used to recover VMs in the target environment. For many organizations, the significant operational flexibility and cost-savings offered by next-generation CDR is worth the trade-off of a slightly longer RTO, especially for certain workloads.
Scenarios that might apply to you
You currently have DR, but it is too expensive You want to reduce costs by moving to “server-less” DR in the cloud that leverages the on-demand provisioning and pay-as-you-go billing capability of the cloud while enabling fast DR site setup and tear down.
You have production on-premises, have no DR, and want to implement it You can’t afford on-premises DR with the OpEx expenditures required to maintain duplicate datacenters, but you need an offsite backup and recovery capability for BC and compliance purposes. Ultimately, you need a low OpEx DR site with minimal-to-zero active management.
You have Production in the Cloud, and want to implement DR You need an “all-in-the-cloud” DR solution because traditional DR solutions do not work in the Cloud. Once a strategic choice is made to move production to the cloud, there are only two options for DR: region-to-region within the same cloud or cloud-to-cloud across two platforms as the source and target.
ZConverter’s all-in-one offering for cloud backup, disaster recovery, migration, and ransomware protection
Let us examine the features that make ZConverter’s CDR on Oracle Cloud a compelling solution for organizations of all sizes that run workloads “all-in-the-cloud”, or in hybrid or multi-cloud environments.
ZConverter’s proprietary ZIA technology and how it works
ZIA is ZConverter’s proprietary source-agnostic image format that is different from other solutions because it leverages the same technology across multiple processes i.e., traditional backup, cloud migration and cloud disaster recovery. ZIA provides the ability to backup, replicate, migrate, and recover operating systems, applications, data, associated IT services, and necessary dependencies, all simultaneously.
ZIA actualizes ZConverter’s claim of being “the most uniform, error-free, and reliable cloud recovery process available in the market today” and enables end-users to overcome the significant challenges in completing simultaneous, error-free, and source-agnostic migrations.
Figure 3 illustrates the function of the ZIA image format
The migration problems that most CDR solutions struggle with arise from their lack of proprietary technology, and a heavy reliance on complicated and error-prone open-source or third-party migration technologies, especially those built into hypervisors. Since ZIA is a file-system-level based technology, it allows server administrators to make hardware-level changes, or complete OS patches and upgrades on source servers that run production application workloads, without requiring new full backups. This capability reduces recovery errors in disaster scenarios and lowers long-term IT management overhead.
Any-to-any source agnostic recovery
ZConverter simplifies the recovery and migration process and eliminates errors by reducing the large number of potential conversion scenarios between source and target environments, from over 1,000 to only 1. The profound simplification and repeatability of the migration process, made possible by ZIA, allows for a reliable recovery process regardless of the conversion scenario.
The term “server-less” implies that there is no requirement to pre-provision duplicate infrastructure resources, such as individual workload VMs and dedicated block volumes, replication servers, conversion servers, block storage for thousands of snapshots, etc. in the target cloud environment. Further cost reductions are delivered by storing the full and incremental backups in low-cost object storage, requiring the backup and replication of only the used portion of the source-server disks and by compressing all the workloads for long-term storage until recovery time. The only infrastructure that needs to be deployed in the Oracle Cloud prior to a ZConverter-based recovery is:
A single, Oracle Cloud Compute E4 shape VM with Ubuntu Linux OS, 2 Cores and 8GB RAM, and 60GB root partition storage.
An immutable object-based repository to store full and incremental backup.
A single, small block volume for processing incremental backups prior to transferring them to object storage for long-term storage.
The significant cost reductions delivered with ZConverter’s truly “server-less” CDR solution allows organizations of all sizes to implement a CDR solution that reduces the compute and storage footprint and takes advantage of the usage-based on-demand provisioning and pay-as-you-go billing capability of the Cloud.
Next-generation backup technology
Conventional file-system-level backup capability is at the center of ZConverter’s claim of being a next-generation CDR solution. The ZIA proprietary file-image format prevents the following backup problems that plague most present-day CDR solutions:
Recovery errors and inconsistencies.
Excessive I/O overhead on the source servers that can cause application service outages.
Runaway storage requirements at the source and target locations.
Long replication times and high data transfer costs.
Source network performance bottlenecks when simultaneously replicating to the cloud and performing local incremental backups.
Lack of ransomware protection.
Replication lag and backlogs that prevent RTO and RPO times from being achieved.
Point-in-time RPO
ZConverter delivers comprehensive Point-in-Time Recovery/RPO capability, allowing recovery down to the folder and file level from backups taken months or even years prior. Some CDR solutions can only recover from the latest incremental backup. Be mindful of solutions that claim exceptionally low RTO or “real-time” replication. These “real-time” solutions often fail to provide comprehensive point-in-time RPO recovery capability.
Low-cost immutable storage
ZConverter CDR supports secure low-cost immutable object-based storage at source and target locations. This adds significant cost-savings and security capabilities for protection from ransomware attacks or inadvertently deleted or corrupted data. In addition to being highly available, scalable, and inexpensive, a key benefit offered by immutable storage is the ability to configure it in a read-only WORM (Write Once Read Many) state, rendering it more secure than block storage.
Ransomware protection
ZConverter delivers ransomware protection via secure immutable object-based storage and comprehensive RPO. Many present-day CDR solutions advertise real-time replication and deliver continuous data protection (“CDP”) and they fail to offer comprehensive RPO by only allowing recovery from the last incremental backup. This is also a weakness of pure HA solutions. In the event of a ransomware attack, applications and data must be recoverable from the point-in-time that they were not encrypted or corrupted. This could require recovery using backups made days, weeks, or even months prior to the attack. ZConverter supports comprehensive RPO, so that even if the ransomware-encrypted data propagates across the recovery infrastructure, there is still a way to successfully recover into Oracle Cloud using a clean and consistent set of operating systems, applications, and data.
Uniform OS recovery and legacy OS support
ZConverter’s ability to recover virtually any workload successfully to the cloud, from any cloud or on-premises environment, is critical to a successful CDR implementation. The uniformity of ZConverter’s recovery process is mandatory for meeting the complex technical challenges faced when migrating multiple different machine types to the cloud simultaneously. This includes ZConverter’s ability to migrate physical server workloads to the cloud as virtual servers and the ability to recover the VM back to a physical server on-premises if needed. ZConverter supports many legacy Operating Systems such as: Windows 2003, Windows 7, CentOS 4.x-5.x, Ubuntu 12.x-14.x, Debian 8.x-9.x and Oracle Linux 5.x-6.x.
Compression and encryption of only the used portion of source disk
ZConverter can replicate only the used portion of each source disk and encrypt and compress all Operating Systems, applications, and data, from the time that they are backed-up at the source location to the time that their files are needed for recovery in the target location.
Scalable replication-processing
ZConverter can scale the replication-processing at the source location without causing application servers to crash, hang, or perform poorly, causing users to experience I/O lag and frozen screens. Other CDR solutions lack this capability because they do not support dedicated replication devices, and instead perform replication-processing on the source servers or their shared storage device. They often use resource-intensive “copy-on-write” snapshot technology which also adds to the replication I/O overhead problem.
ZConverter addresses these challenges by deploying dedicated ZConverter Disaster Recovery Manager (“ZDM”) servers at source and target locations where they offload replication-processing duties from the source servers. Multiple ZDMs can be deployed at the source location to handle the intensive processing needed by heavy, write-intensive scenarios, or scale-out environments with large amounts of servers and data. For extreme replication scenarios, the ZDMs can even be upgraded to high performance multi-processor Bare Metal servers running local SSD drives or can be attached to a high-performance Network-Attached Storage (“NAS”).
ZConverter’s all-in-one solution
Each ZConverter cloud disaster recovery software subscription supports the following capabilities:
Cloud disaster recovery
Enterprise backup (On-premises and cloud)
Ransomware protection
Server migration
The ZConverter ZDM software and ZIA proprietary image format technology is at the heart of the ZConverter source-agnostic next-generation CDR solution.
Real-world use case: VMware on-premises to Oracle cloud disaster recovery
An organization has compute and storage resources that reside on Oracle Cloud, in another Cloud, and on-premises running on VMware, and they want to achieve hybrid-cloud and multi-cloud backup and recovery into Oracle Cloud of all their workloads and infrastructure i.e., operating systems, applications, data, associated IT services, and their dependencies.
Figure 5 illustrates VMware on-premises to Oracle Cloud DR with required ZDM Port Information.
ZConverter Cloud disaster recovery setup and recovery process
Phase I: ZDM setup, backup, and replication
Provision virtual or Bare Metal servers at both your source and target environments that will operate as the ZDM servers.
Load ZDM software and synchronize the source and target ZDM servers.
Install an agent file on all source servers.
Within the ZDM portal do the following:
a. Complete an automated full backup of source servers i.e., a compressed and encrypted backup of operating systems, applications, and data.
b. Replicate backup from source location to target storage repository within Oracle Cloud tenancy.
c. Setup incremental backup and replication schedule(s) within the ZDM portal (e.g., daily, weekly, or monthly).
Phase 2: Disaster recovery
Either manually provision VMs in the Oracle Cloud target cloud environment, or automatically provision them simultaneously using a pre-configured Terraform or Ansible script.
Transfer backup repository data located in Oracle Cloud object-based storage to block storage.
Then, using the ZDM portal, push selected ZIA image files to the VMs from the block storage. Files are automatically decompressed and unencrypted by the ZDM at this time.
Manually access target VMs in Oracle Cloud target cloud environment to verify all services are operating properly.
Once services are verified to be functioning correctly then release server into production.
Figure 6 illustrates the ZDM deployment and the process for Cloud Disaster Recovery.
Note: Prior knowledge of the Linux operating system is very helpful for the ZDM setup.
Value of ZConverter’s cloud disaster recovery solution on OCI
The ZConverter cloud disaster recovery solution on Oracle Cloud offers the following value:
Protects workloads residing in Oracle Cloud using two separate regions as the source and target locations.
Protects workloads residing in other cloud platforms using Oracle Cloud as the target recovery location.
Protects VMware workloads residing on-premises using Oracle Cloud as the target recovery location.
Provides the flexibility to use full and incremental backups, stored at the source and target locations to recover into either their source or target environments.
Provides ransomware protection for workloads, by storing the full backup and successive incremental backups in Oracle Cloud’s low-cost read-only immutable cloud object-based storage and supports comprehensive point-in-time RPO that allows recovery from previous full and incremental backups, all the way down to the file and folder level.
Why cloud disaster recovery on Oracle cloud with ZConverter?
A key DR concern for organizations with on-premises applications is excessive cost of datacenter duplication, whereas their key concern with cloud-based applications is whether they can migrate those applications error-free to a separate heterogenous cloud platform, and then recover them on-demand in a disaster scenario. As with the insurance analogy, if the up-front cost is perceived as too high it makes its purchase unjustifiable. The cost-prohibitive nature of DR is still the main reason organizations do not implement DR solutions. Instead, they continue to rely on only data backups to recover from disasters. This approach has proven to be insufficient to guarantee business continuity and often leads to disastrous outcomes as previously discussed.
ZConverter and Oracle Cloud have solved the problems of excessive cost and lack of CDR reliability. One major advantage offered by Oracle Cloud is its scalable, low-cost object-based storage that is based upon high performance NVMe technology. As a result, Oracle Cloud provides fast data-retrieval times that make object-based storage highly suitable for low RTO on-demand disaster recovery scenarios. Oracle Cloud’s price-performance advantage provides a recovery capability that no other cloud provider can match today. ZConverter-Oracle Cloud DR is the most reliable and cost-effective DR solution for organizations looking to protect their Bare Metal and virtual hybrid-cloud and multi-cloud workloads that are spread across multiple locations.
Summary
The high costs and technical shortfalls of most CDR solutions have prevented widespread enterprise adoption of this promising next-generation capability. Together, Oracle and ZConverter have addressed these shortfalls in a CDR solution that is reliable and inexpensive to deploy and maintain, while meeting all applicable technical criteria. The innovation does not stop there, Oracle Cloud is committed to providing the advanced tools and services that you need to build your entire production and development environment on our world-class Infrastructure-as-a-Service (“IaaS”) platform that leads the industry in price and performance.
* ZConverter is not designed for real-time replication zero-downtime HA or cluster migrations. ZConverter does not support FreeBSD, Solaris, or AIX.
A great way to test-run this next-generation cloud disaster recovery solution is with a free POC demonstration or even a 30-day free trial of Oracle Cloud Infrastructure, that includes US$300 in Oracle Cloud credits and Oracle Cloud Always Free services. A ZConverter cloud disaster recovery software subscription includes free setup, a POC demonstration, and disaster recovery drill training in your own environment. For the POC, you can even choose either your on-premises environment or another cloud platform, or both, as your source locations.
Authors
Ananta Ahluwalia
Principal Cloud Architect
Senior Technology Innovation Catalyst and solutions architect guiding enterprise-scale transformation and strategic change through the innovative use of technology. Delivers high impact, high quality, and mission critical private and public sector projects. Focused on Cloud, On-Premise middleware, and enterprise deployment of Content Management platforms in complex architectures. Skilled in driving high value digital innovation, revenue and profitability through customer engagement and harnessing the power of people.
