The Recurring IT Nightmare
Daniel Alonso, Sr. Consultant III, Caribbean
Peter Shand, Chief Technology Officer, Americas
As the next hurricane season draws closer, people, whether in their businesses or in their personal lives, start thinking about how to prepare for the possibility of disaster. National Geographic reported that 2017 was the most expensive in US history in terms of damage from hurricanes. Additionally, the memories of Harvey, Irma and Maria are still very fresh in the minds of those who live in the Southern United States, the Caribbean, and Central America.
Many organizations often find themselves scrambling to solidify their contingencies and strategies for recovery and response to an emergency in the few months before the start of the hurricane season.
Most businesses in the above-mentioned regions have some sort of disaster preparedness plan, although it may be poorly implemented and/or has never been truly tested. Remedying this situation can require less time and expenditure than you might think, and will allow you to achieve a heightened level of preparedness. It is important to bear in mind that in some instances not putting your plan to the test may lead to a false sense of security as it might be possible that the existing plan may no longer meet the requirements of a changing business and a fresh approach may be needed.
New businesses or those that have established new presence in elevated risk regions can sometimes need to deploy infrastructure simultaneously for primary datacenters as well as recovery sites. It is often very costly and requires many months or years of planning to properly implement a newly created plan. However, properly developing a plan can afford leadership the possibility of taking a holistic look at the business continuity requirements for the organization, and this process may lead to closer alignment between the disaster recovery strategies and the needs of the business. Opting-out of any sort of IT disaster recovery strategy should be significantly less prudent, based on last year’s very active hurricane season. Companies should, at a minimum, consider the adoption of some basic and affordable techniques to protect key resources.
Although disaster recovery entails many technological components, including centralized application delivery and distributed endpoint computing; this article focuses on speedy and successful recovery from disasters that affects the application infrastructure which is essentially, the datacenter. Additionally, these recovery technologies can be divided into two categories:
(1) Ready-To-Go and (2) Work-To-Do
Ready-To-Go technologies allow for application workloads to be in a hot\warm state and can be active and serving users within minutes. Businesses that employ these types of technologies require low RTO\RPO (Recovery Time Objectives and Recovery Point Objectives). Here are some options:
- Built-in software such as vSphere and Hyper-V Replication provide solid protection for virtual workloads.
- Products like Veeam and Zerto add automation and orchestration to the process of initiating an actual failover process or even testing one non-disruptively.
- The public cloud provides additional options natively with solutions from Microsoft Azure and Amazon Web Services. This is important for customers who do not own geographically diverse colocation or want to extend the distance between their sites. Storage array replication can be a major catalyst for enabling disaster resilience.
- Pure storage and Tegile provide replication solutions that can help to accomplish site synchronization however HPE presents a unique approach to a traditional technology.
- 3Par Remote Copy comes ready to replicate out-of-the-box from a hardware perspective over both fibre-channel as well as traditional IP mediums. No additional modules are needed as the controllers come with both port types standard. In a nutshell, all the efficiencies that are gained by implementing a custom ASIC driven storage architecture are preserved through its method of federation. Synchronous, Asynchronous Periodic, and Asynchronous Streaming are all possible out of the box with HPE 3par, which allows for great flexibility depending on budgets and connectivity constraints.
- Businesses that are setting up new infrastructure or at the end of a depreciation cycle for their datacenter storage and compute have an option to leverage HPE Simplivity. Simplivity de-duplicates input/output (I/O) operations which results in huge benefits from an efficiency standpoint. This transfers into its federation\replication as well. Essentially the Simplivity cluster acts as a WAN accelerator in geo-distributed deployments. This allows for significant savings in telecommunications costs.
- The optional rapidDR software suite adds the mechanism to orchestrate a failover plan that is designed to move workloads natively inside Hyper-V or VMware.
The technologies described above may be employed by organizations across the full spectrum of disaster recovery readiness.
Not all businesses require systems to be up and functional within minutes or even a few hours (higher RTO\RPO) and as such may, employ disaster recovery technologies that require a restore process to complete before workloads are able to be powered-on and accessible again. This restore process usually means copying data from the backup media to some available primary storage whether on a schedule or manually at the invocation of a disaster recovery plan. This recovery strategy is based on using secondary storage as the target for copies of application data. Here are some things to consider:
- In most cases it is not possible to run active workloads using secondary storage although Instant Recovery by Veeam allows for some workloads to be started using copies hosted on secondary storage.
- HPE’s StoreOnce Catalyst technology in tandem with Veeam can provide that Instant Recovery functionality. The StoreOnce appliance not only maximizes the capacity once the data has landed on the repository, but also deduplicates data on the source side, preventing any unneeded transmission of data to the appliance and when federating multiple appliances across geographical distances. This results in a massive reduction of data stored and can allow for increased retention times to meet compliance.
- Another recently announced extension of this technology is the ability to store backups on supported cloud storage such as Amazon AWS (S3) via StoreOnce “CloudBank”
- The public cloud is fast becoming a desirable option for hosting replicas and backups of critical application data as AWS, Azure and Google Cloud Platform offer virtual workload backup to cloud solutions, allowing for virtual machines to be spun up in the public cloud in a disaster scenario. This can be a low-cost option for the budget constrained organization.
The practice of business availability encompasses and supports all pillars of the datacenter. Unified Tech’s team of world class experts are ready to help you plan, develop, and implement recovery strategies that support your organization’s vision for availability.