We grill high-end backup kid on its cloudy data protection stake

Datos IO has a unique approach to data protection that is at odds with legacy media-server-based data protection and the use of purpose-built backup appliances. It says its emphasis on how applications store and manage their data makes it different from newer data protection vendors such as Cohesity, Rubrik and Veeam. We …

[tomjoyce64]

Kudos

It's about time somebody tried to break the model for B/U and recovery for cloud. This is one of those product ideas where when the lightbulb goes off and you understand what they are trying to do, it's like wow, smart, that makes sense, and why not? I think they need to a better job of marketing it so it isn't such a journey to get to the light bulb moment, and they need to publish comprehensive benchmarks to prove it works, but other than that they have a great idea here.

[Anonymous Coward]

Looks good on paper

Um, sounds wonderful on paper..... a lot of talk in this article about the well known limitations of the current backup solutions, but little/no substance about how DatasIO is actually doing anything any different, details, examples, case studies, customers, etc... nothing...

[Cragdoo]

"Veeam is based on a media server architecture where data is piped from VMWare (or equivalent servers) to target agents on the media server with attached storage. The deduplication is done partly on the source and partly on the target."

VMWare ...sigh and there is no such thing as a media server in Veeam

Here are the key differences with Veeam:

"All protected data must flow through Veeam media servers and this approach will not scale in a geo-distributed multi-cloud world. The media servers will become a choke point and the use of attached storage is 10 times more expensive in a cloud deployment model."

there it is again, media server ....doesn't exist, so will be difficult to compare.

"Veeam primarily deals with opaque blocks and files and thereby cannot provide fine-grained data protection or advanced data management services such as search."

Not true, Veeam has both application aware processing/explorers, to allow for fine-grained protection, and indexing on VMs for search functionality.

"While Veeam benefits from application transactional consistency from the use of VSS, this cannot be termed application-centric as Veeam does not have any insight into the structure of the data being protected."

Contradiction to the previous satement, now Veeam does have application transactional consistent backups??

[PeterSmails]

Education is an important part of helping organizations make informed decisions, so it’s beneficial to review the facts of each comment:

First comment

"Veeam is based on a media server architecture where data is piped from VMWare (or equivalent servers) to target agents on the media server with attached storage. The deduplication is done partly on the source and partly on the target."

VMWare ...sigh and there is no such thing as a media server in Veeam

Here are the key differences with Veeam:

"All protected data must flow through Veeam media servers and this approach will not scale in a geo-distributed multi-cloud world. The media servers will become a choke point and the use of attached storage is 10 times more expensive in a cloud deployment model."

there it is again, media server ....doesn't exist, so will be difficult to compare.

Datos IO Response

As per the Veeam web site:

https://helpcenter.veeam.com/backup/vsphere/backup_architecture.html

we quote the following text:

"The source host and the backup repository produce two terminal points between which VM data is moved. Backup data is collected, transformed and transferred with the help of Veeam Data Mover Services. Veeam Backup & Replication uses two-service architecture — one Veeam Data Mover Service controls interaction with the source host during data transfer and the other one controls interaction with the backup repository. The Veeam Data Mover Services communicate with each other and maintain a stable connection. All backup infrastructure components engaged in the job make up a data pipe. VM data is moved over this data pipe block by block, so that processing of a single VM includes multiple processing cycles.

When a new backup session starts, Veeam Backup & Replication performs the following actions:

1. Veeam Backup & Replication deploys the runtime process on VM guest OSes via the guest interaction proxy (for Microsoft Windows VMs) or backup server (for VMs with other OSes).

2. The target-side Veeam Data Mover Service obtains the job instructions and communicates with the source-side Veeam Data Mover Service to begin data collection.

3. The source-side Veeam Data Mover Service copies VM data from the snapshot using one of VMware transport modes, as prescribed by the backup proxy server settings. While copying, the source-side Veeam Data Mover Service performs additional processing — it consolidates the content of virtual disks by filtering out zero data blocks and blocks of swap files. During incremental job runs, the Veeam Data Mover Service retrieves only those data blocks that have changed since the previous job run. Copied blocks of data are compressed and moved from the source-side Veeam Data Mover Service to the target-side Data Mover Service.

4. The target-side Veeam Data Mover Service deduplicates similar blocks of data and writes the result to the backup file in the backup repository.”

Based on common industry parlance, the target-side Veeam Data Mover Service is referred to as a media server.

2nd Comment

"Veeam primarily deals with opaque blocks and files and thereby cannot provide fine-grained data protection or advanced data management services such as search."

Not true, Veeam has both application aware processing/explorers, to allow for fine-grained protection, and indexing on VMs for search functionality.

Datos IO Response

Per Veeam’s website, Veeam allows for recovery at a file level using browser functionality:

https://helpcenter.veeam.com/docs/backup/vsphere/data_recovery.html?ver=95,

However, while this might have been sufficient in an earlier generation, today’s RTOs and RPOs demand an application to be up and running from a fine-grained restore without the application user having to browse files and stitch things together.

A similar problem is with the search paradigm:

https://helpcenter.veeam.com/docs/backup/em/understanding_search.html?ver=95.

All Veeam search allows is to index files, which is not aligned with what 99.99% of users expect: search for unstructured text or structured rows and columns using a query pattern on a data source.

3rd Comment

"While Veeam benefits from application transactional consistency from the use of VSS, this cannot be termed application-centric as Veeam does not have any insight into the structure of the data being protected."

Contradiction to the previous satement, now Veeam does have application transactional consistent backups??

Datos IO Response

Datos IO did not make the statement that Veeam does not have application transactional consistent backups. So, we are not clear what statement the poster is referring to.

Always feel free to reach out to [e-mail] or contact us via our website at www.datos.io, we’re happy to discuss further!

[Cragdoo]

You quote Veeam's own text, which at no point uses the phrase media server. If you're going to educate people at least use the correct terminology. The correct term you're looking for is Repository.

"Based on common industry parlance, the target-side Veeam Data Mover Service is referred to as a media server."

Not sure you can call a single service a server. A service, well lots of services actually, typically run on a server.

"However, while this might have been sufficient in an earlier generation, today’s RTOs and RPOs demand an application to be up and running from a fine-grained restore without the application user having to browse files and stitch things together."

Oh you mean like SQL database restores? https://helpcenter.veeam.com/backup/howtosql/how2_sql_restore_with_vesql.html

Restore database to the original or different location, to the state as in the selected restore point of SQL Server VM.

Restore database to the selected point in time.

Restore database to the state prior to undesired transaction table

I'll rescind the final statement, having read it again.

[Max Melville]

I work for Assured Data Protection a Rubrik partner, after reading the comments i felt compelled to put my point of view across.

Tarun: “Both Rubrik and Cohesity (aka, “the PBBA Replacement Vendors”) are based on a media server architecture where data is piped from VMWare (or equivalent servers) to target agents on a hyper-converged media server... this approach will not scale in a geo-distributed multi-cloud world and will become a choke point.”

Comment: Rubrik is not a hyper-converged appliance running backup components. Rubrik manages multi-threaded backups without physical or logical media servers. In fact, Rubrik was first to market with a scale out file system and distributed task scheduler that scales without limit, across hybrid cloud and geographies.

Tarun: “The approach of protecting opaque blocks and files cannot provide fine-grained data protection or advanced data management services such as search. While VSS provides application consistency, both Rubrik and Cohesity lack any insight into the structure of the data being protected.”

Comment: For virtualized and physical systems Rubrik builds a metadata index on ingest, which allows for global, real-time search; granular recovery of objects (i.e., emails, files, folders); and granular recovery of points-in-time (i.e., log-enabled database restores). Rubrik was first to market with real-time search in the 2015 1.0 Release.

Tarun: “The integrated storage offering is an order of magnitude more expensive in a cloud deployment model…Both Rubrik and Cohesity treat cloud data stores as long-term retention targets while Datos IO uses these data stores as secondary storage...

Comment: Rubrik has some of the most exciting integrated cloud capabilities in the market. Rubrik instantiated in the cloud (Cloud Cluster) protects cloud native apps or apps instantiated in the cloud. Cloud Cluster uses cloud block storage (e.g., EC2) for instant recovery, but also uses cloud object storage (e.g., S3) for archive. Users set policies for utilization of each storage type per their business need. Moreover, the recent Alta release enables application migration with data from on-premises VMware to AWS.

I'll happily provide a free no obligation demonstration to prove the above.

[Anonymous Coward]

I have to disagree with this statement: "Rubrik is not a hyper-converged appliance running backup components"

Rubrik and Cohesity are absolutely "hyper converged" backup appliances. They've taken the control server, media server, and storage and merged them into a scale out appliance. That is the definition of hyper-converged. Obviously their marketing doesn't position it that way but this is what they've done.

[Rickatron (Veeam)]

“To clarify the comments made about Veeam, there is no component called a Media Server or a Media Server Architecture. That puts a lot of question into the comparison made by the CEO of Datos IO in this interview. I encourage readers feel free to reach out to myself or anyone at Veeam for more information.”