How a Well-Designed Architecture will Maximize Virtual Resources and Grow with Your Environment - a Close Look at vRanger Pro DPP

The release of vRanger Pro 4 in 2009 included three new letters in the product name, which are DPP. DPP stands for Data Protection Platform, and its introduction is the first platform for data protection which is optimized for image-based data handling.

Being first to market with an image-based data protection platform was not an easy step for Vizioncore. Some might say that we risked losing ground in the market. We did need time to stabilize DPP based on the fact that it was a large change in the architecture of the vRanger Pro product. But we believe that concerns about the time required to create and stabilize this platform is short-term thinking. For the long-term, we took the time in 2009 to design, develop, and introduce a stronger architecture that will prepare our customers today to keep up with the growth of VMs and data long into the future.

Scalability and performance are critical requirements for managing images and data in virtual environments. We've experienced the growth and development of virtualization alongside our customers. We've lived it with them. What we saw coming a year ago is the same challenge that we see happening now: a growth curve of VMs and resulting data unlike any that the industry has faced before.

To help our customers prepare for this challenge, we had to find a way to make images faster to collect, smaller to transmit and store, and simpler to use for all types of data recovery. Being able to support simultaneous backup and recovery jobs, for example, turns out to be a key requirement for keeping up with VM and data growth. Compression being on or off during the backup process to support dedupe appliances is another key example. There are more. Vizioncore developed these capabilities - and built them into DPP - by listening to our customers, working alongside them, and thinking ahead about not just the challenges that we all face today in managing current environments but also about the challenges which are coming.

The result is DPP. DPP is a scalable, efficient foundation for data protection which can be deployed for efficiency today with assurance that it will grow with your VMs and data into the future. With Vizioncore, you won't reach a dead-end in your data management. With Vizioncore, you will not experience the performance bottlenecks - the uh-oh moments - in the use of network bandwidth, storage, and administrator time and attention.

Instead, you can rely on Vizioncore as having the the platform for image-based data management on which you can build for the future. You are future-proofed with DPP and vrRanger Pro 4.5.

Key Architecture Design Features in DPP

Agentless protection - vRanger Pro injects a binary at runtime to manage the backup. This binary works with the VMware snapshot to read data from a VMDK file, compresses and transmits the backup image, and then cleans up behind itself by removing itself from the system. There is a huge different in the administrative work required with this type of design, as compared to an agent-dependent solution: with agent-less, administrators do not need to deploy software on each virtual server system, do not need to track licenses, do not need to upgrade agents to match file system patches and application upgrades. Instead, vRanger Pro queries vCenter to know which version of the binary is required, and injects that appropriate version at runtime.

For more information on how vRanger Pro works with VMware snapshot to take the backup, read this previous blog post.

And, yes, vRanger Pro will support Microsoft Hyper-V in the future. The first part of our data protection platform introduces support for Hyper-V next week at Microsoft TechEd. We look forward to hearing from you about our social media campaign this summer, in which we hope to hear what YOU think of our P2V2P capabilities and what they make possible in your environment.

Service-based architecture - A service-based architecture enables jobs to be submitted and executed to completion, even when components of the vRanger Pro software -- such as the GUI console -- are stopped. This is a critical difference that makes possible multi-stream, concurrent job handling.

Multiple, concurrent backup streams - DPP enables multiple, concurrent VM backups to occur at the same time. Images are not collected sequentially, but in parallel, even on the same physical host and simultaneously across hosts. The result is a dramatic reduction in the time require to create a backup copy of the images on a system. Most environments have many VMs today; even very small environments often have TBs of VM data to manage. The growth of VMs in these environments is very high, which makes this capability critical to ensure that image-based data management remains practical.

Multiple, concurrent restore jobs - If you are depending on image-based protection for Disaster Recovery, when a LUN-ful of VMs gets overwritten, or in any of a number of other real-life scenarios, you need the ability to efficiently restore multiple VMs simultaneously. DPP enables simultaneous restore. vRanger Pro GUI console enables multiple VMs to be selected for restore, without having to walk through the restore wizard separately for each VM. Instead, all VMs required can be selected for restore at the same time. This type of ease-of-administration feature is critical for real-world operations.

Image compression at the source - Compressing images at the source, before they are sent downstream over networks and onto storage is critical for maximizing virtual resources.

Dynamic virtual resource management - DPP is dynamically aware of virtual resources in the environment, and optimizes the use of those resources for executing backup and restore jobs -- without over-running the environment. Environment queries include: vCenter interrogation of VMs and ESX and vSphere hosts, network storage protocols being used, and how storage is shared between VMs. Administrators can set parameters on the use of available resources, to control their use.In this way, DPP eliminates the complexity of manually orchestrating these activities and eliminates the need for administrators to constantly monitor and intervene to ensure jobs complete successfully.

This type of dynamic resource management is especially critical in a virtual environment. In a physical environment, administrators know which applications are running on each server, application peak hours of operation, and how much of the system resource is available to support backup, and when. In a virtual environment, the operation of each individual VM can affect the entire virtual server and all other VMs on the system. Capacity can be consumed quickly, in terms of I/O and CPU loads, and jobs can compete and collide. DPP ensures that systems continue to operate while multiple protection jobs execute and complete.

Direct-to-Target architecture - DPP enables transmission of data from virtual servers, where it is collected, directly to target storage devices which are available on the network. The key advantage of this architecture, is the seamless scalability that it provides for adding and expanding storage in the environment. This architecture also eliminates the need to send backup data through an intermediary server - often called a Media Server - which becomes a performance bottleneck.

Archive files created and managed independently - DPP creates and stores many archive files independently, which enables additional parallel processing and performance gains. This is also key for ensuring that the backup archive is designed to have integrity, ease-of-administration, and efficient portability.

Archive files stored with intuitive, VM-specific names - Storing backup archive files with VM-specific names makes it far easier for administrators to browse, find, and restore the exact VMs that they need. Otherwise, administrators must keep tract of which archive file includes which VM(s) - and must communicate this information among administrators.

VM archives independently stored and portable to other media - Storing separate files for each VM save point makes them independently portable to thumb drives, and most importantly, to tape. For subsequent sweep-to-tape operations, only the changed files must be moved to tape - which ensures that tape use is efficient and optimized, along with the time required to perform the job, and restore later.

Archive files remain unmodified after creation - A key requirement for archive management, is to ensure that backup files stored in the archive remain unmodified after they have been written. Keeping these files intact and unmodified ensures their integrity for recovery. This is also key in some environments, for meeting compliance and corporate governance oversite requirements.

Synthetic, single-pass restore from multiple archives - DPP enables efficient restore, by intelligently reading only the relevant portions of the archive files. When restore of a VM or a file requires blocks from multiple archive images, such as across a full and multiple incrementals, then DPP enables vRanger Pro to build an in-memory bitmap of the archives and reads only the additional blocks required for the restore.

DPP is the Better Foundation for All Data Protection from Vizioncore/Quest

DPP is not just the foundation that enables scalability and performance from vRanger Pro, but is rather the broader foundation for all future data protection capabilities from Vizioncore/Quest. DPP is the integration point for vRanger Pro and vReplicator, and other capabilities in the data protection portfolio. As platform capabilities are expanded and enhanced - including capabilities like job reporting, scheduling, encryption and others - they will be immediately available to benefit all data protection capabilities using the platform.

As shown in the illustration, vReplicator integration with vRanger Pro and DPP is under way and will shortly benefit from the DPP scalability and performance capabilities built into its architecture.

To learn more about the DPP architecture, read the Architecture Overview attached to this blog post.

Anonymous