How To Choose The Right Tiering Solution
| File tiering is new to most organizations. Given this, few people are experts in the technology. In fact, many of us are unsure of even what questions to ask. It turns out that there are several very important questions. A few features can make the difference between long-term success, or a nightmare that gets worse every year.The purpose of this paper is to show you the questions that matter and which answers are the right ones. |
EXECUTIVE SUMMARY
“Always focus on the front windshield and not the rear view mirror.”
– Colin Powell
- While many of us have a depth of experience in managing primary storage, tiering is new to most of us
- This can mean that we may not even know the right questions to ask, let alone what their answers should be
- A few of the key questions are common sense, but the answers cannot be assumed, such as “Does the system capture all the elements of my files?” (There are plenty of tiering systems that do not.)
- Other questions require out-of-the-box thinking, like “Will I be able to get at my files after your company is gone?”
- The rest of the issues surround what it means to live with your chosen solution. Tiering is forever. Among other things, this means that flexibility is key, as your needs and priorities are certain to change over time
- Make the right choices and you will be happy for years. Make the wrong choice, and the cost of fixing it can be enormous
NEVER ASSUME…
You might assume that every backup product, every tiering product, makes a complete and correct copy of the files it touches… You can make that assumption, but you would be wrong. There are several backup products and tiering products that discard elements of your files, and even more that discard metadata.
In some cases the product is just bad. But a more common scenario is that the product is derived from a Linux-based solution that now needs to work with Windows files. Windows files have elements (such as alternate streams) and metadata that Linux files do not have, and which Linux doesn’t support. This means that a Linux-based product cannot simply “copy” Windows files. It must copy the body of the file and some of the metadata as a Linux file, and then create a secondary object to contain the elements of the Windows file that the Linux file system doesn’t support.
Unfortunately, this is rarely done. And, just as few people read backup logs, even fewer know how to check that a file – any file, Windows or Linux – has been copied correctly. All that most people see, and what most people check, are the body data and the visible metadata.
This means that alternate streams, enhanced security, and extended metadata may be lost. Alternate streams are used by both Windows and Macs, enhanced security is used by Windows 7 and later Microsoft platforms, and extended metadata involves file permissions that Windows has that cannot be reproduced in Linux.
Criteria #1: Do you care whether all of your data reaches Tier 2, or is it all right for the tiering system to throw some of it away?
THE CATALOG CONTROLS YOUR FUTURE
Five years ago you worked on that project… what was its name…? And those files, the ones that you need… you called them what, again?
The files that you tier today may or may not have stubs. But someday, for sure, they won’t. (A NAS full of stubs is still a full NAS. Best practice is no stubs, or stubs that last only a year or two after a file is tiered.)
When the stub is gone, how does an end-user find their tiered files? A good tiering system will have an end-user access portal that retains the file’s original security. This portal allows end-users to search the metadata of the archived files that they have (legitimate, secure) access to just as they can with their files that are currently online.
Once they locate what they need, then the portal allows them to download the results to their desktop.
Meanwhile, a good system will have a second portal that allows you, the system administrator, to search and find whatever you may need across all accounts.
Finally, an open (non-proprietary) catalog also means that you can run whatever reports you need to understand what’s being tiered over time and how your money is being spent.
Criteria #2: Can your end-users service themselves once there is no stub, or will they need you to do the work for them? Independent of the users, do you have what you need to do your job? Can you get what you need from Tier 2 cost-effectively?
TECHNOLOGIES COME AND GO…
You don’t want your files going with them. The protocols and formats used by NASes are standard and open. CIFS, NFS have been around for years and files move easily from one place to the next.
However, there are many tiering solutions which, just like backup products, have a proprietary back-end. Today, there are Fortune 100 companies with rooms full of obsolete hardware that they have to retain in the hope of being able to read their legacy backups. There is no need to put yourself in this position with tiering.
A good tiering solution not only has an open catalog, it has an open back-end. If the solution is discontinued in the marketplace, if its vendor fails, you will still have complete access to your files.
Criteria #3: Are you willing to bet that a proprietary solution will outlive your data?
THE ONLY CONSTANT IS CHANGE
Those of us who are old enough remember a day before the “Cloud”, a time when the file server race was between Microsoft and Netware and when the only purpose-built NAS came from a company called Network Appliance. The only thing we know for sure about the future is that it will be different from today.
What this means for tiering is that you need flexibility. Single-purpose, point-to-point, “I only move files from source X to destination Y” solutions are unlikely to survive the test of time. And, as previously discussed, the cost of changing from a closed system to a different system can be enormous – more than the first system cost.
A good tiering system will accommodate multiple sources and destinations, so that you are not locked in to any one vendor’s platform technology. It will allow you to create rules that are granular in scope so that you can manage different classes of files and different communities separately.
Criteria #4: Does the system empower you, or limit you?
WHAT IF…?
As this is being written, there are rumblings in Congress that if the Industry doesn’t do a better job of securing data, then the Government is going to legislate. A company doing business in all 50 states in the US is subject to more than 200 data protection laws – most of which are honored in the breach, if at all.
Beyond simple compliance, which is regularly ignored, there can be subpoenas and legal holds, audit requirements, etc… A good Tier 2 system interfaces with the applications you may need to satisfy these requirements and others.
Criteria #5: If compliance becomes an issue, can your tiering solution do what will be required?
IN SUMMATION
Despite the five questions, there is really only one issue that matters. Once your files have been tiered, will you still have the data elements and flexibility needed to solve a problem you can’t anticipate today. Let’s review the key questions whose real purpose is to ensure that you have everything you need both the day after tiering and 10 years later:
- Do you care whether all of your data reaches Tier 2, or is it all right for the tiering system to throw some of it away?
- Can your end-users service themselves once there is no stub, or will they need you to do the work for them? Independent of the users, do you have what you need to do your job? Can you get what you need from Tier 2 cost-effectively?
- Are you willing to bet that a proprietary solution will outlive your data?
- Does the system empower you, or limit you?
- If compliance becomes an issue, can your tiering solution do what will be required?
Don’t assume that every tiering solution on the market will address all of these concerns.
But do understand that once you implement a tiering solution, there’s no easy way to turn back. Moving 100s of terabytes or petabytes of files takes months and months. A simple oversight today could mean catastrophe in the future. Your organization needs the highest level of service possible. Deliver on that promise by choosing a solution that gives you the flexibility you need, and empowers you and your end users to succeed.
You can find the full document here: Choosing The Right Tiering Solution
