Protecting data at remote offices and distributed data centers is challenging to say the least. Typically, organizations try to implement backup locally by installing a backup server and some disk or tape. IT then tries to manage the process remotely but if the backup software or hardware fails then IT needs “borrow” one of the location’s personnel to help. One of the big challenges facing IT is how to get all this data back to the primary data center, which acts as the hub.
The Hub and Spoke Dilemma
Many organizations want to, at some point, have all their data in one location, typically the primary data center. For those organizations with remote offices and data centers, this means somehow devising a strategy to backup data locally at the remote office and then somehow ship it to the primary data center. That “shipment” can be either physical or digital. If it is physical then the organization has to have software that can copy the disk-based backup to a tape device and send that tape to the primary data center. Then somehow integrate that tape into the primary data center’s backup catalog.
The problem with physical backup movement is there is an overall resistance to tape within IT. There is also a legitimate concern about the time it will take to make the tape copy, box that tape up and ship it to the primary data center. Then once that tape arrives at the primary data center it has to be unboxed, inserted into a library and potentially reindexed into the backup infrastructure. The overarching problem is that the first part of these steps requires a human, who likely has no IT experience, to perform them.
An alternative is to replicate the backup data on disk directly to the primary data center. Thanks to technologies like deduplication, it is possible to replicate only net new or changed data to the primary data center and either hardware or software can drive this process. But it is still necessary to integrate the remote offices’ backup data into the primary data center’s backup catalogs.
The problem is the transfer is very dependent on the quality of the network segment as well as the distance from the primary site. Additionally, those remote site backups still require cataloging by the primary data center’s backup solution.
In both cases, IT is hoping that the application that they use in the primary data center will also be usable in the remote offices. The remote office backup solutions must be easy to manage remotely, easy for on-premises personnel to use and have an effective means to transfer data. Lastly, it is necessary to either send or replicate all the data at the on-premises location, both its data and the data from the remote offices, to a disaster recovery site. For remote office, the DR requirement results in a third copy of data.
Is The Cloud a Better Hub?
Organizations need an alternative to using the primary data center as the centralized hub. The cloud may be the ideal hub. In a cloud model, IT sends primary data center and remote office data to a public cloud provider, which acts as the centralized repository. Data is copied once and there is one primary store of all backup data. Some solutions will cache data at each remote office and the primary data center so that restores of recently protected data can be quickly serviced but the actual movement of data is just one step. A DR copy can automatically be created by replicating the cloud copy within the cloud infrastructure.
The other advantage of using the cloud as the hub is that it almost guarantees remote management will be of high quality since all sites essentially require remote management. It also means that traveling or vacationing IT personnel can remotely manage the primary data center data protection process.
Using the cloud also better positions the organization to adhere to various legal and regulatory standards that dictate where data can reside since the larger public cloud providers have multiple data centers in multiple regions.
There is nothing wrong with the hub-spoke method for protecting remote offices; it’s just that the primary data center may not be the best candidate for protection given the availability of such a strong alternative. By centralizing both primary and remote office backups in the cloud, the organization creates a single pool of protection and automatically replicates it to an alternate data center.
To learn more about protecting data in distributed enterprises join us for our on demand webinar “Overcoming the Five Distributed Data Protection Challenges”.