ExaGrid Detailed Product Description - DS 10
ExaGrid Detailed Product Description - DS 10
ExaGrid Detailed Product Description - DS 10
ExaGrid sits behind the existing backup server and replaces straight disk,
inline deduplication appliances or tape backup storage, both onsite and offsite.
ExaGrid appliances are comprised of processors, memory, networking, IPMI, RAID 6, and a hot spare drive per
appliance, using enterprise-class SAS drives and ExaGrid software. See the ExaGrid Technical Specifications
data sheet for more details.
Each appliance plugs in and is virtualized into a shared system with a single user interface, global
deduplication, and automatic load balancing. The media server is connected to the same network and sees
the appliances as one or more NAS shares, Veeam Data Mover targets, Veritas NetBackup OST targets, or
S3 Object Storage. Since each appliance includes the appropriate amount of processor, memory, Landing
Zone disk, deduplicated repository disk, and bandwidth for the rated data size, performance increases as more
appliances are added to the system.
ExaGrid’s patented zone-level technology stores only the changed data at a granular level from backup to
backup instead of storing full copies. ExaGrid uses zone stamps and similarity detection.
This unique approach reduces the disk space required by an average of 20:1 and from 10:1 up to 50:1
depending on data type, retention and backup rotation delivering unparalleled performance for the fastest
backups and restores.
ExaGrid is the only vendor that allows backup applications such as Veeam and Commvault to keep their
deduplication turned on and then ExaGrid further deduplicates the data for increased storage efficiency.
ExaGrid supports automated job management utilizing Veeam SOBR, Veritas NetBackup Single Target
Storage Pool, Commvault Spill & Fill, Oracle RMAN Channels, HYCU Scale-out and other automated job
management facilities. ExaGrid globally deduplicates data across all appliances in a scale-out system.
Global deduplication ensures that all data is deduplicated regardless of the number of appliances in a
system. In addition, global deduplication allows organizations the flexibility to redirect backup jobs to any
appliance at any time while maintaining data deduplication globally across the entire system.
Since over 90% of restores and 100% of instant VM recoveries and tape copies are done from the most
recent backup, this approach avoids the overhead incurred from “re-hydrating” data during critical restores.
As a result, restore, recovery, and copy times from an ExaGrid system are an order of magnitude faster than
solutions that only store deduplicated data. The most recent data is available in an undeduplicated native
backup format for fast restores to keep user productivity up by quickly recovering deleted files, overwritten
files, corrupted files, encrypted files, etc. The non-network-facing Repository Tier stores the most recent
backups as well as all the long-term retention in a storage efficient deduplicated format. Legal discoveries, SEC
audits, financial audits, regulatory audits etc. have a lead time to recover the data where recovering for users
does not. ExaGrid has found the balance between fast user data restores and long-term retention data storage
efficiency.
Restore Performance
To ensure fast performance, the backup storage files system must be able to handle large backup jobs unlike
NAS storage which optimized for files or SSD storage which is optimized for transactional databases. In
addition, to further increase performance, advanced backup protocols, concurrency and other techniques
need to be deployed. Standard disk or SSD storage is not optimized for backup.
ExaGrid is designed and optimized for large backup jobs. ExaGrid uses all the following techniques to achieve
the fastest backup performance. In side-by-side tests ExaGrid is faster than any other storage for ingest
performance resulting in the shortest backup window.
y Backups are written direct to the disk Landing Zone (no inline deduplication)
y ExaGrid is integrated with backup applications scale-out functionality that allow backup
applications to write to any ExaGrid appliance that is not still processing data allowing all
appliances to be utilized for performance. The combination of the Veeam SOBR, Veritas
NetBackup Single Target Disk Pool, Commvault Spill & Fill, Oracle RMAN Channel, HYCU
Scale-out, and others allows for front-end performance load balancing. Data can go to
any appliance at any time as ExaGrid supports a scale-out storage architecture, global
deduplication and automatic load balancing.
y File system optimized for large backup files. Most storage is optimized for files or database
transactions. Backup files are very large and require a file system optimized for backup.
Since ExaGrid is dedicated backup storage it is optimized for large backup jobs
y Use of advanced protocols. ExaGrid has full appliances in a scale-out system. ExaGrid runs
software that brings functionality but also integrates with backup applications. ExaGrid
can gain an additional 30% ingest performance over CIFS and NFS by using protocols or
layers built specifically for large backup jobs such as the Veeam Data Mover and Veritas
NetBackup OST.
y Job concurrency allows for backup jobs to be written in parallel allowing full use of all
appliances, resulting in improved performance.
When tested against primary storage or SSD storage ExaGrid is typically 2X the performance due to SSD
file system limitations. When tested against inline deduplication appliances (which are compute bound by
performing the deduplication before writing to the disk), ExaGrid is typically 3X to 4X faster. ExaGrid is as
much as 5X faster than general purpose files systems such as Microsoft ReFS to disk.
In contrast, ExaGrid’s scale-out approach with global deduplication adds full appliances—including memory,
processor, and bandwidth, as well as disk. The figure below shows the differences between how the two
different architectures cope with data growth over time.
ExaGrid appliances include scalable computing software, which allows them to virtualize and share data
storage capacity with one another (automatic load balancing across all repositories). This scalable system
(shown above) can expand as data grows by adding appliances, providing up to: 13.82PB raw capacity, 12PB
usable capacity, and allowing full backups of up to 6PB in a single scale-out system. Multiple systems can be
used at a single location, and up to 16 separate locations can be managed through a single user interface.
ExaGrid supports backup data from multiple sources, including a variety of backup applications and
database dump utilities. Performing deduplication in the backup application software limits the ability to
have all data from all sources stored and deduplicated in a single target device. Unless 100% of your backup
data passes through that particular backup application, a Tiered Backup Storage appliance such as ExaGrid’s
is the best choice to meet the requirements of your entire environment.
In contrast, backup application software solutions that have incorporated deduplication by definition
only support their own backup application, with its own media server software and its own backup client
agents. These solutions are not able to support backup data from other backup applications or utilities.
For example, if you have a physical environment, for backup applications that employ data deduplication
but want to use a separate utility for VMware (ex: Veeam) and also do direct database dumps, only the
data running through the physical system’s backup application will be deduplicated. Also, deduplication
in backup software ranges from 2:1 to 5:1 and therefore uses a lot more disk and bandwidth as retention
grows versus target-side deduplication appliances that employ far more aggressive deduplication
algorithms, since with appliances, resources are dedicated to the task. Dedicated appliances achieve ratios
on average of 20:1.
ExaGrid allows for Veeam and Commvault to leave deduplication turned on and ExaGrid will further
deduplicate the data for increase storage efficiency.
Veritas’ Open Storage Technology is another popular feature that allows for more integrated offsite
data protection, and it is important to check whether these features are supported if you are using
Veritas NetBackup, OST allows for faster performance, better management, and unbalanced onsite and
offsite retention.
ExaGrid is the only solution in the market for Veritas NetBackup Accelerator that can reconstitute a full backup
in its Landing Zone so that a complete backup is ready for restore in its already hydrated form for faster
restores and VM boots.
ExaGrid also supports Veeam Fast Clone, which allows for 30x faster synthetic fulls, which takes minutes.
Automatic resynthesis of the synthetic fulls into actual full backups takes place in parallel with backups, which
allows for the fastest restores & VM boots for Veeam data in the industry.
S3
M365
B&R
S3 for M365
Veeam Data Mover for B&R
Veeam’s Scale-Out Backup Repository (SOBR) allows backup administrators using Veeam to direct all jobs to
a single repository made up of ExaGrid shares across multiple ExaGrid appliances with global deduplication
in a scale-out system, automating job management to ExaGrid appliances. ExaGrid’s support of SOBR also
automates the addition of appliances into an ExaGrid system as data grows by simply adding appliances to
a Veeam repository group. The combination of Veeam SOBR and ExaGrid’s appliances in a scale-out system
creates a tightly integrated end-to-end backup solution that allows backup administrators to leverage the
advantages of scale-out in both the backup application as well as the backup storage. The combination of
Veeam backups to the ExaGrid disk-cache Landing Zone, the integrated ExaGrid-Veeam Accelerated Data
Mover, and ExaGrid’s support of Veeam SOBR is the most tightly integrated solution on the market for a scale-
out backup application to scale-out backup storage.
ExaGrid can also allow users to turn Commvault deduplication off, with Commvault compression either
turned on or off, to increase backup performance while retaining the same cost storage as leaving
Commvault deduplication on with ExaGrid’s additional deduplication impact.
ExaGrid supports Commvault Spill & Fill for automatic job management where all jobs are sent to ExaGrid
appliances in the system by Commvault automatically. Jobs can be sent to any appliance at any time as
ExaGrid has both global deduplication across all appliances in the system and automatic load balancing of
all long-term retention data repositories.
1. A single Oracle database can be up to 6PB in size and can be backed up in parallel to a
single ExaGrid scale-out system.
2. The database backup performance is accelerated as the sections are backed up in parallel
across multiple appliances in a scale-out system.
3. The database backup performance is maximized as each new section is automatically sent
to the highest performance availability NAS shares and/or appliance, resulting in the best
possible performance based on NAS share and appliance ingest availability.
4. If any appliance fails, the segments are automatically redirected to the active appliance,
providing for automatic failover.
5. The most recent database is stored in an undeduplicated form in the ExaGrid disk-cache
Landing Zone, allowing for fast restores while still allowing for storage efficiency as all
long-term retention data is stored in deduplicated form. This avoids the lengthy data re-
hydration process of inline scale-up appliances that only store deduplicated data.
6. As the database data grows, the backup window stays fixed in length as full appliances are
added into a scale-out system bringing compute with capacity. This eliminates the forklift
upgrades associated with inline scale-up deduplication appliances.
ExaGrid’s architecture and implementation have multiple facets of reliability and redundancy, allowing
organizations that are considering disk-based backup appliances to make informed vendor selections.
ExaGrid offers the following ease of use, redundancy and security features, some of which are
explained below:
y Single user interface for all appliances in a system and across sites
y RAID 6 protection with a hot swappable spare
y Redundant hot swappable power supplies
y Active Directory for management interface and backup target security
y SNMP and sys-logging interface for integration with enterprise management apps
y Role-based access control
y Retention Time-Lock – ransomware recovery
y Two-factor authentication
y Data encrypted at rest
y Data encryption while replicating over the WAN
y Security checklist makes it easy to apply best practices
y Data is checksummed to ensure data integrity
y Internal self-describing database
ExaGrid provides Tiered Backup Storage with a front-end disk-cache Landing Zone and separate Repository
Tier containing all retention data. Data is written directly to the “network facing” ExaGrid disk-cache
Landing Zone. Then it is tiered into a “non-network-facing” long-term retention Repository Tier (tiered air
gap) where it is stored as deduplicated data objects to reduce the storage cost of long-term retention data.
As data is tiered to the Repository Tier, it is deduplicated and stored in a series of objects and metadata. As
with other object storage systems, the ExaGrid objects and metadata never change allowing only for the
creation of new objects or deletion of old objects when retention is reached.
ExaGrid’s approach to ransomware allows organizations to set up a time-lock period that governs the
processing of any delete requests in the Repository Tier. In addition, this tier is non-network-facing and not
accessible to hackers. The combination of a non-network-facing tier, a delayed deletion for a period of time
and objects that never change (immutable) are the elements of the ExaGrid Retention Time-Lock solution.
For example, if the time lock period for the Repository Tier is set to 10 days, then when delete requests
are sent to the ExaGrid from a backup application that has been compromised or from a hacked CIFS or
other communications protocols, the data in the Repository Tier is time-locked for up to 10 days against
any deletion. The data in the Landing Zone will be deleted or encrypted, however, the Repository Tier data
is not deleted upon an external request for the configured period of time. When a ransomware attack is
identified, simply put the ExaGrid system into a new recover mode and then restore any and all backup
data to primary storage. The time lock period is separate and in addition to the days, week, months and
year or retention that is set by the backup application and stored by ExaGrid in the Repository Tier.
The solution provides a retention lock, but only for an adjustable period of time as it delays the deletes.
ExaGrid chose not to implement Retention Time-Lock forever because the cost of the storage would be
unmanageable. ExaGrid already has the long term backup retention so it would be redundant to have a
separate store with retention lock. With the ExaGrid delayed delete approach, all that is needed is up to
an additional 10% more repository storage to hold the delay for the deletes. ExaGrid allows the delay of
deletes to be changed from the 10 day default.
1) Data is deleted in the ExaGrid disk-cache Landing Zone via the backup application or by hacking
the communication protocol. Since the Repository Tier data has a delayed delete time lock, the
objects are still intact and available to restore. When the ransomware event is detected, simply put
the ExaGrid in a new recover mode and restore. You have as much time to detect the ransomware
attack as the time lock was set for on the ExaGrid. If you had the time lock set for 10 days, then
you have 10 days to detect the ransomware attack and put the ExaGrid system in the new recover
mode for restoring data.
Note: see ExaGrid’s Retention Time-Lock data sheet for more detailed information.
During normal operation, the RAID controller does consistency checking of the data on its disks in the
background, correcting any disk media errors using the parity disks.
The ExaGrid software continually scrubs the repository data, confirming checksums and automatically
repairing any deduplicated data that does not match its checksum using data from remote site(s). This
automatic repair of deduplicated data is covered by one of ExaGrid’s patents.
Logging Filesystem
Backup data is kept in the ExaGrid internal storage on an industry-standard logging filesystem where file
activity is logged for integrity and quick repair after an unclean shutdown.
Data can be encrypted during replication between ExaGrid systems. Encryption occurs on the sending
ExaGrid system, is encrypted as it traverses the WAN, and is decrypted at the target ExaGrid system.
This eliminates the need for a VPN to perform encryption across the WAN.
Backing up your data to an ExaGrid appliance at your primary site dramatically reduces the amount of disk
space required to store all of that data due to its high-performance data deduplication capability. In a multi-
site ExaGrid environment, the onsite ExaGrid system is only sending deduplicated data—the backup data that
changes at a granular level between each backup—over the wide area network (WAN) to the offsite ExaGrid
appliance. The offsite ExaGrid appliance is ready for data restore and fast recovery in the event of a disaster or
other primary site outage.
If the replication is one way only, the second site/offsite ExaGrid can be half the capacity of the primary site
ExaGrid greatly reducing overall cost.
Replication between ExaGrid systems across a WAN can be scheduled for the day of the week and multiple times
throughout each day. Each scheduled period allows for bandwidth throttling which limits replication to only
use the assigned bandwidth. The combination of scheduling flexibility and bandwidth throttling allows for the
maximum efficiency of WAN bandwidth used for replication. Replicated data can be encrypted over the WAN
using a customer’s VPN or by utilizing the ExaGrid built-in replication encryption.
Private Cloud
y Replicating to an ExaGrid at a customer’s second data center (DR site)
y Replicating to an ExaGrid at a third-party hosted data center (DR site)
Hybrid Cloud
y Replicating to a Managed Service Provider (MSP)
Public Cloud
y Replicating to an ExaGrid VM in a public cloud (Amazon AWS, Microsoft Azure), where
DR data is stored in the public cloud and billed by the GB per month using OPEX budget
Multi Hop
ExaGrid redefines the economics of backup by helping you contain costs at every point in the life cycle —
up front and as data grows over time.
When comparing ExaGrid appliances deduplication in the backup application software, it is important to keep
in mind that using deduplication in the backup application software typically requires greater resources on
the backup server—more processing power, more memory, and more disk. Software deduplication merely
shifts the backup performance bottleneck to the media server. Using data deduplication in the backup
software uses more disk and bandwidth over time and does not allow for backup environment flexibility such
as using a separate utility for virtualized backup, direct TAR backups, and direct database dumps such as SQL
dumps or Oracle RMAN dumps. ExaGrid’s performance will be the fastest and deduplication will be three
to ten times more efficient. In addition, ExaGrid allows Veeam and Commvault deduplication to be turned
on and ExaGrid with further deduplicate that data greatly increasing the deduplication ratio to save on
storage costs.
Other appliances that use inline, block-level deduplication do not support a scale-out architecture and are
therefore more costly to scale. Instead of adding capacity by adding full servers, only disk shelves are added
over time as data grows. But, at some point, the single front-end controller becomes a bottleneck due to its
fixed processor, memory and bandwidth resources and can no longer handle the backup load. Eventually,
the entire front-end server must be replaced with the next higher capacity unit in a “forklift upgrade.” In fact,
you may have to spend as much for the front-end controller upgrade as you originally spent on the original
system, including disk shelves. In addition, all data is always deduplicated. For each restore, recovery, and copy
request, the data has to be put back together, or “re-hydrated,” which can take hours to days.
In addition, unlike other appliances that “end-of-life” in as little as 18 months and are incompatible with
newer models from the same vendor, ExaGrid’s scale-out architecture allows you to “mix and match”
different capacities and generations of appliances within a single system. Only ExaGrid protects your backup
investment from obsolescence.
Repository Tier
y Low-cost long-term deduplicated retention storage
y Industry-leading 20:1 data deduplication
- Global Deduplication
y Adaptive Deduplication
y Deduplicates and replicates during the backup window
y Strong offsite RTO and RPO
y Retention Time-Lock for Ransomware Recovery
y Non-network-facing tier
y Delayed deletes
y Immutable deduplication objects
Scale-out Architecture
y High density 2U appliance models for rack space efficiency
y Scales to a 6PB full back up in a single system
y Fixed-length backup window as backup data grows
y Eliminates forklift upgrades of scale-up architectures
y Mix and match appliances – any age and any size
y No planned product obsolescence (no end of life of maintenance and support)
y 7 different capacity sized appliance models
y Scales as your data grows
Programs
y Product price protection for 5 years
y Maintenance and support price protection – won’t go up more than 3% per year
About ExaGrid
ExaGrid provides Tiered Backup Storage with a unique disk-cache Landing Zone, long-term retention
repository, and scale-out architecture. ExaGrid’s Landing Zone provides for the fastest backups, restores,
and instant VM recoveries. The Repository Tier offers the lowest cost for long-term retention.
ExaGrid’s scale-out architecture includes full appliances and ensures a fixed-length backup window as data
grows, eliminating expensive forklift upgrades and planned product obsolescence. ExaGrid offers the only
two-tiered backup storage approach with a non-network-facing tier (tiered air gap), delayed deletes, and
immutable objects to recover from ransomware attacks.
Visit us at exagrid.com or connect with us on LinkedIn. See what our customers have to say about their
own ExaGrid experiences and why they now spend significantly less time on backup in our customer
success stories.
ExaGrid reserves the right to change specifications or other product information without notice. ExaGrid and the
ExaGrid logo are trademarks of ExaGrid Systems, Inc. All other trademarks are the property of their respective holders.
©2024 ExaGrid Systems, Inc. All rights reserved.