Bringing things together is a good thing.
Here at Arcserve, we’ve been in the business of providing our customers a full spectrum of data protection options for years: From traditional methods of tape and disk backup, to quick-as-lightning full system snapshots, bare metal recovery, and onward to live replication and high availability across both LANs and WANs. We have fully embraced a variety of virtualization and cloud platforms, too. People come to us for the variety of options we provide, so they can truly customize a solution to fit their environment’s needs.
With the advent of Arcserve Unified Data Protection, we’ve brought these proven technologies together under essentially one console – and made them even better by adding new technologies. Unified Data Protection (UDP) is able to capture full-system snapshots of physical or virtual nodes, replicate them within a LAN or across a WAN, create virtual clones of the nodes, and even create tape archives of it all.
That’s fantastic, but, I’m often asked by customers … how?
My answer is: The heart of UDP is the Recovery Point Server. It’s the place where everything begins.
It’s not just a place to put backups – think of it as the coordination point for nearly all the operations within UDP. Including:
- RPS saves recovery points (snapshots) in its Data Stores
- RPS helps in the creation of Virtual Standby instances – that is, virtual clones which can be brought online (locally or remotely) from multiple recovery points
- RPS acts as the hub for Global Deduplication across the entire enterprise
- RPS Data Stores allow for unique encryption and compression settings for each store
- RPS is the source and target for replication of the recovery points
- RPS controls the bandwidth needed to move recovery points from site to site
- RPS is the exit and entry point for Data Seeding during implementation
- RPS becomes the data source for migrating snapshots to tape
- RPS can replicate recovery points to an RPS installed into the cloud
See what I mean? It’s not a one-trick-pony.
Implementing an RPS
A Recovery Point Server is a software application installed on either a physical or virtual node. (And soon, we’ll have an option to purchase a pre-configured appliance!) It stores backups from multiple sources, either from UDP agents on physical machines, or agentless snapshots from virtual targets. It can also be deployed as an all-in-one with the Manager UI, or in a distributed deployment.
Within each RPS, we define one or many Data Stores. Each Data Store has its own definitions of where the backup data is stored, deduplicated, compressed and encrypted. Multiple data stores per RPS server allow total control over how you store your data. Need to separate HR information from your Customer Accounts? Create separate data stores for each, and encrypt them differently. Have multiple customers who replicate their snapshots to you? Separate the data stores for each. And you can still manage all this from the same console.
From there, create a data protection plan that includes a variety of actions. First, back up the data. Then… the world’s your oyster.
- Replicate remote sites’ data from their own RPS back to your corporate datacenter to ensure local recoverability but also full system restorability in the event of catastrophic loss
- Create Virtual Standby copies of your source machines so they can be powered on – onsite or off – at a moment’s notice, from any point in time we have snapshots
- Use the Arcserve Backup integration to move Recovery Points off to tape, with full granular restorability
- Create an RPS in the cloud, and send Recovery Points there for archive purposes
- A combination of… all of the above.
The UDP management console can manage multiple RPSs in an organization, each one holding multiple data stores with their own unique compression, encryption and deduplication configurations.
Speaking of deduplication…
Global Deduplication – storage and network savings, even better than ever
So you might be asking yourself … How much storage is needed for all these additional recovery points? Great question.
Deduplication is the process of identifying unique content within a backup set, and storing it only once – while allowing for quick and easy restoration of files, folders, application data and even complete servers. While Arcserve Backup has had server-side deduplication for years, UDP has a brand new Source-Side approach that’s possible with the help of the Recovery Point Server. The result is a totally new way of protecting your data.
There are two types of deduplication within UDP:
Source-side Data Deduplication
Ensures that only unique data from the agent is sent to a recovery point server. We take the snapshot, split the data into unique blocks, and then send a quick I/O index to the RPS to determine if those blocks are already present there. The RPS replies with a list of blocks that aren’t already stored within it. Then – and only then – does the source actually send the data. The beauty of this method? Significantly less bandwidth is used than server-side deduplication, because no redundant content is streaming across the network.
Ensures that only unique data from multiple agents are backed up to a recovery point server. If similar data blocks are present on multiple nodes, only one copy is backed up to the recovery point server. For example, if two source nodes have the same OS, application, or content, all the redundant content is compared and deduplicated at the RPS server.
This approach is also applied between multiple RPS servers across the LAN or WAN – ensuring that each RPS has only one copy of each unique block. Ultimately, Global Deduplication saves bandwidth, storage, time and money.
So you see, the Recovery Point Server enables so many functions ~ it truly acts as the heart of UDP. And in conjunction with the RPS, Global Deduplication allows you to decrease network traffic and storage requirements dramatically.
Ultimately, combining these new Arcserve technologies, UDP saves bandwidth, storage, time and money.
Now that’s bringing it all together!