Arraya Insights | October 3, 2016
A recent study sought to gauge the level of confidence organizations have in their approach to data protection – and the results were not pretty. The Dell EMC study revealed less-than-overwhelming confidence in here-and-now recovery abilities and painted an even bleaker picture of future prospects. Considering this study pinned the average cost of data loss at more than $760,000 and the average cost of unplanned downtime at $570,000, it’s certainly in companies’ best interest to do what it takes to feel better about data protection.
When asked to assess their efforts in Dell EMC’s Global Data Protection Index, organizations had the following to say:
- 63% are less than “very confident” in their ability to meet service level agreements around fully recovering systems and data – both onsite and off
- 81% believe their current data protection solutions won’t be able to handle all future business challenges
Putting together the data protection puzzle
Speed, reliability, and scalability are three essential components of any data protection strategy – and can all be found in Dell EMC’s Avamar/Data Domain solutions. These formerly separate solutions have become more tightly integrated in recent years, forming a pairing that can help solve data protection challenges for organizations of all sizes.
Avamar’s role in backup and recovery has changed quite a bit in the last three years or so. This solution used to be a network storage and data protection RAIN device. However, lately it’s progressed from this role as a landing spot up to the brains of the operation. Avamar, in most scenarios, is used solely as the backup DMA and no longer is leveraged as the target for the backups themselves.
Meanwhile Data Domain, which started its life in the data center as a purpose-built NAS device capable of high rates of data ingestion and inline deduplication, has stepped into the role of backup target in an integrated solution with Avamar. This is due in part to its deduplication algorithm’s effectiveness with common data types, higher ingest bandwidth, and smaller, denser footprint.
Together, these products combine for a dynamic backup and recovery solution providing new features such as instant recovery for virtual machines and low bandwidth replication. In the market for a new backup and disaster recovery option? Looking for a partner to help you through the process? Here’s what a typical Arraya-deployed Avamar/Data Domain solution looks like – as well as why our approach might be exactly what you need.
Arraya’s approach to Avamar/Data Domain deployments
Right off the bat, while this may be a “typical” Arraya deployment, it’s not the only option available. We always work with customers to come up with backup and disaster recovery strategies and architectures that address their needs – now and down the road.
Arraya recommends deploying Avamar either as a virtual appliance or as a single Avamar node with minimal storage (2TB) capacity on the device itself. Traditionally, in an integrated solution, the Avamar node capacity is reserved for meta data only. In fact, we recommend configuring the integrated solution to force all backup data be sent to Data Domain.
In addition to the single Avamar node, our sizing usually includes a midrange Data Domain model with sufficient capacity to encompass a three or five-year growth cycle based on data analysis done by the assigned Solutions Architect. The Data Domain 2500 model is the most common unit used which comes with (6) 1Gbe NICs onboard. Our implementation engineer will recommend configuring two of these NICs into a failover (active/passive) bond and then binding the Data Domain host name to the bond’s assigned IP. The result provides a redundant network path to the Data Domain over this interface. This is used for command/control and replication and requires no external network configuration to support. The functionality is supported by the Data Domain itself.
What about the other four NICs on the 2500 you say? Arraya’s engineer will configure individual IPs on each of them and then roll all four into an interface group which can be accessed by all clients. When backup is initiated, the client will reach out to Data Domain to gain access to an IP it can use during the process. Data Domain will consider the workloads each IP is running at that moment and assign the appropriate IP to maintain network load balance. This functionality also provides fault tolerance. In the event a link goes down, the hand shake between client and Data Domain will occur again (over the command/control path) and Data Domain will assign a new IP link to use to continue the backup uninterrupted. Like PowerPath for backup!
Building your ideal data protection environment
This deployment methodology allows for the marrying of two seemingly disparate technologies to work in tandem to provide a robust BRS solution. From cache files used by the Avamar binaries to build lists of unmodified files that don’t need to be processed, to Distributed Segment Processing done by Data Domain binaries (commonly known as DDboost), the solution reduces necessary network bandwidth significantly. Unique data that does need to be sent is processed by a varying length segment algorithm that identifies patterns in the data stream that are not unique and only stores them once. Summary vectors are created from these segments and built in memory so the Data Domain can quickly determine whether it’s seen a pattern before.
An oversimplified example would be if you were to back up a library the word “the” would occur possibly millions of times. As those books are read and sent to Data Domain, the pattern of 1’s and 0’s for the word “the” will only be stored once. Each subsequent appearance after that isn’t stored or sent across the network. Instead a reference to the original instance is made.
The same technology that speeds up the backup and reduces network bandwidth is used in reverse to speed up the restoration process for a great deal of data types. Both Avamar and Data Domain employ methodologies to ensure that the data sent to them remains reliable and is protected against corruption. So you can sleep easy at night knowing your backups are safe and sound and there to be restored should the need arise.
Arraya’s BRS solution doesn’t just provide fast, efficiently-stored, and reliable backups, it can scale to meet almost any need. Potentially addressing the 81% of companies who don’t see their current data protection solutions as a long-term answer. IPs, connections, bandwidth – they can all be added without the business incurring true, costly disruption.
The next steps
Partnering with Arraya for your backup and disaster recovery refresh means the above deployment type and features are exactly what you’ll get – if they’re what your business wants. Our award-winning data management practice has no interest in shoe-horning a business into an approach that isn’t the right fit. Instead, the strategy and equipment we propose can be adjusted depending on need.
Ready to get started? Visit: www.arrayasolutions.com/contact-us/ to begin your conversation with our team today. We’re also reachable via social media, including Twitter, LinkedIn, and Facebook. Remember to follow us on each to keep up with our latest blogs, industry insights, and special events.