Shielding Data Assets: Exploring Data Protection and Disaster Recovery Strategies
- 1. CHOOSING THE RIGHT © 2023 StoneFly, Inc. | All rights reserved. Data Protection & Disaster Recovery Solution
- 2. CONTENTS 1. Summary 2. What are Disaster Recovery and Data Protection? 3. What are RPO and RTO? 4. Storage Attributes to Look for in a Primary Storage 5. Building Blocks of a Storage Solution 6. Replication 7. Cloning 8. Snapshot 9. Choosing the Right Storage for Your Needs 10. DR / DP in Modern Datacenter 11. Scalability and Flexibility 12. Data Efficiency 13. Ecosystem Integration 14. Automation 15. Control 16. Create a Successful DR / DP Strategy Using Stonefly’s DR365 site in a box
- 3. The available technologies for disaster recovery (DR) and data protection (DP) are rapidly developing. It is of a paramount importance to understand your storage options, so you can build a strategy that maximizes both flex- ibility and control. Means of disaster recovery and data protection (DR / DP) for the enterprise are constantly changing in response to growing data volumes, faster recovery needs and rising business demands. IT teams must make smart choices to take advantage of the latest innovations and ensure they can keep their businesses online all the time. Contrary to the old more or less passive role, today’s Storage Systems are becoming key players in DR / DP prac- tices. If you’re in the market for new storage, paying attention to a few key factors will help you address your DR / DP challenges. Any storage solution you choose must include replication, snapshot and cloning. Compression and inline dedu- plication are also crucial in helping you conserve both network bandwidth and storage capacity. These technolo- gies must operate at the right granularity to enhance efficiency and control DR / DP costs. Primary storage must provide the tools you need to monitor and control DR / DP processes and the elasticity to utilize a variety of secondary storage targets. You may need to be able to recover and replicate to and from a cloud service provider or public cloud, storage from the same vendor, or secondary storage such as a backup ap- pliance (virtual or physical). Your storage must integrate with the DR / DP processes and tools you currently use or plan to use in the future. This guide will help you understand the storage features that are crucial to modern DR / DP practices and pro- vides detailed guidelines on what to look for in a DR / DP storage solution. Summary
- 4. Disaster recovery is the process of restoring a failed service, application or the operations of an entire datacenter to full operational status following a disaster. This includes natural or human-caused disasters. Having a disaster recovery setup in place is essential for ensuring business continuity. Recently, there have been a number of high-profile outages in the industry that are making the enterprise rethink DR planning: • A power outage caused by storm at the Amazon Web Services (AWS) cloud in Sydney causing up to 10 hours of downtime to a number of large websites and online services. • A Delta Airlines datacenter outage in august 2016 grounded over 2000 flights over 3 days and cost $150 mil- lion. Methods of disaster recovery include replication (synchronous or asynchronous) to a secondary site and recovery from offsite tape. Organizations choose synchronous-replication to protect their most mission-critical applica- tions. What are Disaster Recovery and Data Protection? Data protection, also known as backup and restore, is the process of protecting data from Loss, Corruption or accidental deletion. Example of where data protection comes in to play include a user who accidentally deletes an important direc- tory or file, hardware failures that result in minor data losses, or coding errors that cause a database table to become unreadable. Conventional methods of data protection include asynchronous replication and tape backup. Disaster Recovery Disaster Protection
- 5. RPO (Recovery Point Objective) is a service level that specifies the target point in time to which a service or an application can be recovered. A recovery point objective of one hour means that you will be able to restore all your data for a service or an application in no more than one hour. What are RPO and RTO? RTO (Recovery Time Objective) is a service level that specifies the time needed to restore a service or an applica- tion to operation after a disaster takes it offline. For instance, a recovery time objective of two hours is a commit- ment to have a service or an application back to online status within two hours of disaster occurrence. RPO (Recovery Point Objective) RTO (Recovery Time Objective) The players and the landscape of the DR / DP field are changing rapidly. Keep your options open so you don’t get left behind. In the past, DR / DP processes relied almost exclusively on tape media. However, managing tapes was complicat- ed and as the volumes of data grew, restore times became longer. Applications could end up being down days if not weeks. To accelerate restores, shrink backup sizes and improve service levels, IT teams turned to disk based solutions for protecting their critical data and applications. Over time, superior backup appliances that use com- pression and deduplication technologies to lower storage capacity requirement and costs and enhance efficien- cy became extremely popular in backup and disaster recovery strategies. These days, it’s a common practice to have multiple copies of data on different types of media: DR copies in remote locations for disaster, tape or/and cloud copies for longer term retention, disk based copies for fast recov- ery, and so on. There are also other several important trends affecting the DR / DP landscape as well, such as: • IT teams are being asked to shrink RTO and RPO even further at a reasonable price. • Deduplication and Compression – the enabling technologies that led to the backup appliance, are now widely available in modern storage systems. • The use of snapshots as a first line of defense has become increasingly accepted. The Contemporary DR / DP Landscape:
- 6. Storage Attributes to Look for in a Primary Storage The primary storage you deploy in your corporate datacenters and other locations can make a huge difference to the success of your DR / DP efforts. Here are the Storage attributes to look out for: DATA EFFICIENCY ECOSYSTEM INTEGRATION AUTOMATION CONTROL SCALABILITY & FLEXIBILITY CORE BUILDING BLOCKS + The storage solution must include replication, cloning and snapshots. These technologies make the foun- dation of your DR / DP plan. With DR / DP landscape in flow, it’s important to choose storage solution that leaves your future options open and can scale. Deduplication, compression and other efficiency features are essential technologies to save WAN band- width and storage capacities. Your storage solution should inte- grate with all DR / DP elements including other data, hardware, hypervisors, protection software, etc. All DR / DP functions must be easily automated to enable DevOPs and private cloud, save time and reduce errors. Visibility of all the storage functions and the DR / DP aspects can elimi- nate unwelcome surprises.
- 7. A Storage solution with built-in replication allows you to make copies of important applications, data and VMs regularly, providing multiple points of recovery and protecting against disaster. There are two types of replication: • Synchronous • Asynchronous Synchronous replication guarantees that all data is written synchronously to two separate locations. Even if one replica is destroyed in a disaster, no data is lost. Synchronous replication may affect the rate at which data can be written.Synchronous replication typically has a distance limit to minimize the effects of latency on write perfor- mance. Asynchronous replication is usually based on periodic snapshots. Once an initial data copy is created, only the changes that have occurred since the last snapshot are replicated, saving time and bandwidth. This significantly lowers the burden on production storage systems and networks when compared to full data copies. Asynchro- nous replication may be able to support a variety of bidirectional topologies including one-to-one, one-to-many, and many-to-one, giving you more flexibility in how and where you want to protect your data. Both types of replication (depending on need) can be performed across a campus, within a datacenter, across greater distances, or within a metropolitan area. The greater the separation, the greater the latency and the greater the bandwidth costs. Building Blocks of a Storage Solution Replication Several storage capabilities are essential in enterprise data centers. A Storage solution should not be considered if it lacks these features.
- 8. Cloning gives you the ability to create a writable or virtual copy of a storage object such as a VM or LUN without requiring a full copy. Cloning is extremely useful for data management functions such as disaster recovery, data protection, and others. A failed VM can be quickly restored, in either a primary or secondary location, by cloning a backup snapshot. Use of cloning enables you to do DR testing without disrupting ongoing replication. Cloning also simplifies many other functions including a variety of development and DevOPs tasks such as making a test environment identi- cal to your production environment, creating multiple copies of test data sets or creating identical work environ- ments for developers. Clone Building Blocks of a Storage Solution (Cont..) Several storage capabilities are essential in enterprise data centers. A Storage solution should not be considered if it lacks these features. A space-efficient and fast snapshot is your first line of defense in data protection. In many storage setups, snap- shots are the foundation of most data management functions and data protection. Only recently, has this tech- nology been taken seriously as a first line-of-defense ahead of other data protection solutions. It is worth mentioning that there are significant differences in the implementation of snapshots between ven- dors. Some implementations may still rely on time-consuming data copies. That’s not what you want. New implementations track changed blocks. Some may use copy-on write or COW algorithms that will decrease storage performance after a snapshot has been created. The most efficient implementations use a redi- rect-on-write or ROW algorithm. Snapshot
- 9. Choosing the Right Storage for Your Needs DR / DP in Modern Datacenter As you consider these features, it’s worth paying close attention to the differences between implementations. The most important fac- tors include: Performance Does use of the feature have a big impact on the storage system performance either during or after the operation? Especially for pro- duction storage. Time Delay How quickly does the operation execute? If it fails, do you get noti- fied? Granularity At what granularity can the operation be performed? Space efficiency How efficiently does the operation use storage space? Greater efficiency means less cost. VM, vDisk, LUN, volume, file? If your operations are largely virtualized, you’ll likely be happier with storage that can operate directly at the VM level. In a modern datacenter, the IT team typically takes snapshots of im- portant data and VMs on primary storage as a first line of defense every hour or two. This provides many potential recovery points versus daily backups. For some data, snapshots are enough. More important VMs and data are replicated to one or more secondary locations that may be: cloud, DR Datacenter, or service provider at regular intervals. These secondary copies serve both as backups and provide disaster recovery. Should a failure occur, and depending on the severity of the outage, the IT team can: • Restart the affected services or applications at the secondary location. • Restore necessary data or VMs from a secondary copy. • Restore necessary data or VMs locally from a snapshot. Clones of VMs at the secondary location can be used for testing of DR processes.
- 10. Scalability and Flexibility With the evolution occurring in DR / DP landscape, it’s crucial to choose a storage that gives you the flexibility to update your strategy to meet business needs, minimize costs and support scalability. The key factor in ensuring flexibility is to keep your options open in terms of where you will store your DR / DP copies of your critical data. There are three main options: A backup appliance / other secondary storage device: If you already have such a solution in place, your storage must be able to work with it. And you want to be sure your storage integrates without making the environment too heterogeneous. Sourcing a secondary storage solution from a different vendor will likely require third-party soft- ware. Secondary Storage from the same vendor: Using primary and sec- ondary storage from the same vendor has several potential advan- tages. It can simplify management, since you don’t have to manage different storage in the secondary role. A vendor’s native replication may eliminate the need for third-party software and could deliver maximum efficiency. Cloud Storage: Businesses today are looking at ways to utilize the cloud to reduce disaster recovery and backup costs. The following features should be considered when you’re looking for storage solution with maximum flexibility benefits: Vendor limitations: Make sure to understand any limitations in a vendor’s DR / DP tools. For example, how much flexibility do you have in choosing a secondary device? You may choose all-flash stor- age for your primary storage needs, but it will likely be more cost-ef- fective to replicate to hybrid storage or disk-based as the secondary target. Cloud: Be sure to find out which public cloud vendors the storage can integrate with and find out how extensively each storage choice is used by the cloud service providers. It can be advantageous to choose a service provider that uses the same storage you’ll deploy in your datacenter. Synchronous Replication: Even though you may not currently be planning to use synchronous replication, it may be wise to choose a vendor that offers it as an option. Broad support: Make sure the vendor you choose has broad support for the backup appliance / device, secondary storage and cloud stor- age options. See the “Ecosystem Integration” section for more related information.
- 11. Data Efficiency Choosing the most efficient storage for your requirements can significantly decrease your CapEx and OpEx costs—not just in the WAN bandwidth expenses, but also in the amount of storage capacity purchased. Saving WAN Bandwidth Whatever storage solution you choose should include efficient repli- cation at the right granularity. If you are moving or planning to move to a container model, choose a storage solution that can operate at that granularity. There are two ways replication can save WAN band- width and reduce costs: • By decreasing the volume of data that must be sent. For instance, you may want to replicate a single VM or set of VMs, however, LUN-based storage requires that you replicate an entire LUN with many VMs—or re-architect your storage design. • Data is always sent de-duplicated and compressed over the WAN. Saving Storage Capacity Modern storage systems, whether for primary or secondary storage, should include a range of storage efficiency technologies to reduce the overall capacity you’ll need: Deduplication: This technology can reduce the capacity needed for many common data sets. If your environment is virtualized, depend- ing on how it is configured it may contain a large amount of duplica- tion in the form of identical operating system and application files across many VMs. Backup data often has a very high degree of dupli- cation, though this can depend on the backup methods you’re using. Compression: Data compression also does a good job of reducing capacity requirements for most data types. Compression can reduce requirements for database storage where levels of data duplication are typically low and deduplication may not be much help. Since the effectiveness of compression is highly dependent on the data type, look for algorithms optimized for colder backup data. Thin provisioning: Reduces the amount of storage you need to allo- cate to services and applications upfront, increasing storage utiliza- tion.
- 12. Efficient Disaster Recovery as a Service Offerings If you’re a cloud service provider, maximizing resource utilization while controlling costs is critical. To offer Disas- ter Recovery as a Service (DRaaS), however, you may have to over provision storage to satisfy customer needs in case of a failure. And you may have to provision storage systems for each customer or offer different levels of ser- vice. Modern storage systems with per VM quality of service (QoS) can solve this problem. Per-VM QoS allows you to establish tiers of service on a single array; setting maximum or minimum QoS thresholds for groups or even individual VMs. Now you can consolidate data from multiple customers onto a single array and guarantee that each customer will receive the QoS they are contracted for. DRaaS is a fast-growing and increasingly important use case for cloud service providers, but you need storage that can provide the policy management and SLAs necessary to manage these advanced services. DRaaS offerings allow companies to eliminate the complexity and expense of maintaining a datacenter for DR and shift CapEx to OpEx.
- 13. Ecosystem Integration Because storage is a vital part of DR / DP success, your storage should integrate with all the other elements of your environment includ- ing other data, hardware, hypervisors, protection software, etc. Choose a storage solution with wide integration from a company that understands data protection and can support your business. Make sure the storage you choose integrates with the other elements of your DR / DP and datacenter environments: Your storage solution needs to integrate with the hypervisors you run: • Microsoft Hyper-V, including ODX support. • VMware vSphere, including VASA and VAAI support. +
- 14. Automation In order to meet your DR / DP requirements now and in the future, you will want storage that is easily automated. This will facilitate DR / DP processes and testing, offload IT staff, and enable you to better support DevOps needs and private cloud. IT teams recognize automation as a way to streamline tasks and eliminate or reduce the chance of operator error. This is certainly the case with DR / DP operations - most of which are highly repeatable. You want to be able to automate regularly scheduled backup and replication activities, as well as ad-hoc functions. For example, it’s common to create a snapshot of a VM before upgrading software or installing a patch. Then if something goes wrong, you can immedi- ately roll back to the last good working status. You will also want to be able to automate DR testing processes as much as possible. IT teams are being asked to provide private cloud services through self-service portals with infrastructure resources such as VMs and storage capacity. Storage should provide easy automation to make it possible to include DR / DP choices through the self-service portal. For instance, a user might want to provision a VM to run a new appli- cation and have snapshots of that VM created every one or two hours with nightly replication. Many enterprises IT teams are moving to a DevOps model that bridg- es the gap between internal software developers and IT teams to deliver software features more quickly and with higher quality. If your company is moving towards DevOps, you’ll need to be able to incorporate replication, clones and snapshots as part of automated testing and workflows. Things to look for in Storage Automation Easy orchestration: Storage infrastructure that’s easy to orchestrate can be a big advantage. Once again, it comes down to a question of operating at the right granularity. In a virtualized environment, VM-level operations will simplify the automation process. The same is true for container-level operations, if your company is moving in that direction. REST APIs: REST APIs are the key to storage automation in the cloud. To automate all storage functions including monitoring, repli- cation, clones and snapshots, look for storage that offers a full set of REST APIs. Deeper Integration: Some vendors may offer specific integrations on top of REST APIs to make it simpler to automate storage using popu- lar tools such as VMware vRealize Orchestrator, Python, PowerShell, and various Microsoft tools.
- 15. DR PLAN TESTING Bringing up a DR site to make sure that everything is working proper- ly inevitably interrupts ongoing replication processes. That’s why most organizations find it extremely difficult to effectively test a DR plan. Cloning solves this problem. You can clone everything you need for your DR testing to create writable point-in-time copies. Replication processes can continue to write to the active volumes while you carry out your DR testing using the clones. And since additional storage space is only consumed as changes are made, the whole process is extremely storage-efficient. IT teams need to be able to simulate failures and test DR plans with- out impacting ongoing production workloads.
- 16. Control Your Storage should give you the tools you need to stay on top of your DR / DP aspects and storage functions. In a dynamic IT environment, things are continuously changing. The right storage should give you the control tools you need to make sure your DR / DP processes are working as expected and stay on top of your DR / DP environment. Here are a couple of things you might need to do: • Monitor all primary and secondary storage to ensure they have the performance and capacity to keep up with the DR / DP processes and production workloads. • Monitor all DR / DP processes to make sure they are running as expected. • Identify data that is not being protected. • Troubleshoot any problem that arises. The right control tools can greatly improve your ability to implement a successful.DR / DP strategy and save your team substantial amount of time and trouble.
- 17. Create a Successful DR / DP Strategy Using Stonefly’s DR365 site in a box Choosing the right vendor while keeping the above mentioned guidelines in mind, is instrumental for choosing successful DR/DP Strategy. StoneFly’s DR365 site in a box is a total backup solution for all of your physical/virtual servers and workstations in a single. It is cost effective and offers the following major functionalities for your enterprise. DATA EFFICIENCY ECOSYSTEM INTEGRATION AUTOMATION CONTROL SCALABILITY & FLEXIBILITY CORE BUILDING BLOCKS The storage solution replication, cloning and snapshot features oper- ate at the same level of granularity as your operations. Stonefly’s storage solution provides you much flexibility and scalability to grow your DR / DP plan in the future. It offers diverse secondary storage solutions including cloud storage, backup appliances and storage from the same vendor. Deduplication, compression and thin provisioning can significantly reduce the primary and secondary storage capacity needs and reduce WAN bandwidth requirements. That sums up to significant cost savings. Stonefly’s storage solution integrates with other aspects of your DR / DP environment including your existing secondary storage, backup and DR software, orchestration platforms, hypervisors etc. DR365 automates DR testing pro- cesses and DR / DP functions and also incorporates automated DR / DP functions in your DevOps practices and private cloud offering. Predictive analytics simplify capacity planning. DR365 site in a box pro- vides tools that simplify manage- ment and monitoring of all DR / DP processes. This will make sure you are getting the data protection right and allow you to quickly fix any problems that may arise as you map from the storage to the VM to the application.
- 18. Create a Successful DR / DP Strategy Using Stonefly’s DR365 site in a box (Cont..) DR365 is a purpose-built hyperconverged backup and disaster recovery appliance. The integrated high-performance SAN and embedded EXSi or Hyper-V hypervisor combined with data protection and services make it ideal for today’s challenging environments. It is powerful, easy to use and affordable Storage with all the modern features the enterprise needs for DR / DP for all their data and applications ensuring business continuity and high availability. Furthermore it has Recovery Point and Time objectives of less than 15 minutes for all data and applications. Another distinguishing feature of DR365 is that it offers cloud connect by replicating to cool storage such a Microsoft Azure cloud and also to cold storage such as google and AWS cloud. Organizations including; US Department of Defense, National Institute of Health, United States Army Research Laboratory, Ecker Enterprises, LifeCare Assurance and others have said “No to LUNs & legacy Storage”. With StoneFly they manage only virtual machines, in a fraction of the footprint and at far lower cost than traditional storage. For more information, visit www.Stonefly.com or www.iscsi.com & follow us on Twitter @StoneFlyInc
- 19. www.stonefly.com 2865, 2869 and 2879 Grove Way, Castro Valley, CA 94546 USA. +1 (510) 265-1616