Product Analysis: Scalable NAS for the real world: Redefining NAS Storage

NAS systems have come a long way since SMB & NFS were invented in 1984, and the Dell Fluid File System is proof of that. If one hasn’t looked at a modern day NAS system, this would be a good place to start. Anyone familiar with managing NAS systems over the last few decades is aware of the challenges they can present, and this paper will discuss a few of them. It will also explain how the Dell Fluid File System addresses each of these challenges.

The Scalability Dilemma

One of the biggest concerns for those managing NAS filers is lack of scalability without forklift upgrades. The obvious scalability issue is that most NAS appliances have a finite capacity that is a lot smaller than many environments’ needs. If a customer fills up that appliance, the only choice is to buy another appliance and create another storage island.

A less obvious scalability issue when purchasing a typical filer is that one must decide upfront how much compute and storage capacity the system will need to support at any time in its life. Appliances aimed at supporting only a few terabytes will most likely not be able to support hundreds of terabytes, and those designed to support a few hundred users will be crippled by a load of tens of thousands of users. This requires customers to buy today what they think they need tomorrow. Of course that growth may or may not come, and it may come sooner or later than expected. Predicting NAS requirements 4 years into the future is about as accurate as predicting the weather.

The best case scenario is that the IT team perfectly predicts the I/O load and storage capacity eventually needed by a given application and buys the appliance model capable of that load. Of course, this means that the company will be buying tomorrow’s compute and storage capacity with today’s money. Even if they buy additional capacity only as they need it, they are still forced to buy the more expensive appliance that is capable of supporting that extra storage — which is especially bad if they won’t need that storage for a few years. The best case scenario is that they buy extra computing capacity long before it is needed.

The worst case, and unfortunately the most realistic scenario is that the team responsible for “guessing” future storage and compute needs gets it wrong. If they guess too high, they will buy a much more expensive appliance than they will ever need. If they guess too low, they will eventually be forced to replace the appliance with a more expensive model and then move the disk trays or migrate the data from the smaller model to the bigger model. The smaller model is often discarded as a sunk expense, and the appliance swap creates unwanted downtime. The impact of the worst case scenario is wasted money and unplanned downtime.

Redefining Scalability

The Dell Fluid File System solves these scalability problems with a scale-out approach, which means that they have a system of clustered nodes with a single namespace that can be grown as customers’ needs change. In version 5, Dell expanded their global namespace support to have a single namespace even across multiple clusters. This allows customers to buy the number of appliances that they need to meet their storage and compute needs today, and add compute and storage as they need it — no wasted money and no downtime required. In addition, individual volumes can scale from a few terabytes to four petabytes, also without downtime. Multiple volumes can be combined into a single namespace up to 20 petabytes. Volumes are also thin provisioned, which means that a 20TB volume with 1 TB of data will only consume 1 TB of space. Again this means that customers buy and use only what they need today, and then add to the system as their needs grow — without forklift upgrades and downtime.

Data Management Dilemma

Another issue with traditional NAS systems comes from the kinds of issues that are created when you have hundreds or thousands of users storing whatever they want on their SMB or NFS share. Besides wasting space, allowing users to store data that might not be properly licensed can actually create liabilities for the company, as it can be demonstrated the company actively supported unlicensed file sharing. Depending on the types of data that users store, they can also create a significant amount of duplicate data, wasting more disk space. There are also the issues around governance and compliance, and whether or not an organization can prove who accessed what file.

Redefining Data Management

These issues are also solved with the Dell Fluid File System, as customers can decide on a directory basis what kinds of files are allowed or prohibited. Keep unlicensed music and movies off the system by excluding the appropriate file types (e.g. MP3, MP4, etc.). Duplicate data is handled with a post-process deduplication system that is also configurable per directory, and one of the things that can be configured is when the process will take place. By default, frequently accessed data is not deduped until it is no longer being accessed, but customers can configure this process to happen sooner — or never — based on the needs of a particular dataset. Dell addressed the compliance issues in version 5 with the Change Auditor that can show exactly who accessed each file and how they accessed it.

The Data Protection Dilemma

Another issue with scalability is how data protection will be done. While many NAS systems support replicated snapshots, these snapshots are typically done using the inefficient copy-on-write mechanism that requires three I/O operations when a block required for a historical snapshot needs to be overwritten. The more snapshots are taken and the longer they are kept, the greater impact they will have on performance. This makes copy-on-write inappropriate for data protection of large volumes. Another issue with typical mirrored snapshots is that they require the same amount of storage on both the source and target destination.

Redefining Data Protection

The Dell Fluid File System addresses these challenges as well by using redirect-on-write snapshots, which only require one I/O operation to update a block. This makes them much more scalable for large volumes like what this product offers. They also now support having a longer snapshot history on the destination than the source. This allows customers to keep onsite only what they need for operational recovery, and move older snapshots to a different system.

StorageSwiss Take

In conclusion, the Dell Fluid File System solves a number of challenges with traditional NAS storage. It allows customers to buy only what they need for today’s needs, and add storage and compute capacity as their needs change. Customers can also restrict data types and select certain data to be deduplicated and compressed. Finally, redirect-on-write snapshots allow customers to meet their recovery needs without using traditional backups. It’s a strong offering that is worthy of consideration.

W. Curtis Preston (aka Mr. Backup) is an expert in backup & recovery systems; a space he has been working in since 1993. He has written three books on the subject, Backup & Recovery, Using SANs and NAS, and Unix Backup & Recovery. Mr. Preston is a writer and has spoken at hundreds of seminars and conferences around the world. Preston’s mission is to arm today’s IT managers with truly unbiased information about today’s storage industry and its products.

Tagged with: , , , , ,
Posted in Product Analysis

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 22,221 other followers

Blog Stats
%d bloggers like this: