By Yuval Shavit, Features Writer
As companies try to control the vast amount of important -- and often legally significant -- information they generate daily, many are struggling with a difficult balancing act. On the one hand, they'd like to keep anything that might be important, even years down the line; on the other hand, storing and managing all that data can be a big strain on an IT department. Archiving -- especially email archiving -- is one of the key solutions to that problem, but deploying an email archiving system requires combining technical understanding with your client's legal needs and corporate culture.
In this installment of out Hot Spot Tutorial on email archiving, we'll give an overview of the technical aspects of email archiving. In the final installment, we'll show you how to work with your clients to ensure the solution fits their needs.
Email archiving means creating an index of each message, so that it can easily be found later, and then storing the message itself. Roughly speaking, there are three approaches a company can take in deploying an email archiving system: indexing and storing messages on-site; duplicating each message and sending them to a third party for indexing and storing; or a hybrid approach, indexing messages on-site but then sending the data to a third party for storing.
All three methods require some on-site hardware, and most companies prefer to keep as much on-site as possible for security reasons, said George Crump, founder and senior analyst at consulting firm Storage Switzerland LLC in Fort Worth, Texas. Any email that's sent to a third party will be encrypted, of course, but if your client hires a company to do the indexing as well as storage, that company has to have access to your client's unencrypted messages. That puts your client at some risk, if the third party is hacked or has a problem with a disgruntled employee. That risk is mitigated with the hybrid approach, since the third party never has access to your client's unencrypted data; it just stores encrypted bytes, which your client's in-house indexing and searching software can then request and decrypt as needed.
Of course, rather than deploying an email archiving system, some companies just outsource the entire email server to an off-site third party, which may then implement its own archiving as part of its hosted offering.
The software that intercepts and indexes messages tends to fall into one of two categories, Crump said. The first consists of applications specifically written for email archiving, while the second is backup software that's extended into email archiving -- although that tends to be less well-integrated, Crump said. But generally speaking, email archiving has been around long enough that the programs for it are all good and cost about the same, said Dick Benton, principal consultant at GlassHouse Technologies Inc. in Framingham, Mass.
The other crucial element, if your company is deploying an email archiving system on-site, is the medium it stores messages on. It's important when you pick a target to consider factors such as how quickly you'll need to retrieve the data and how important it is to prove that it hasn't been tampered with. If your client is looking at archiving primarily as a way of putting items that are old and infrequently accessed aside, its focus should be on speed. If it's looking to keep data for legal or regulatory reasons, however, it's important that the medium be "write once, read many" (WORM). It's also very important that the medium can scale to a near-infinite capacity, Crump said, since the amount of data companies create is only increasing.
Many companies also prefer to keep their archives on two kinds of media, Crump said. For instance, if a company is archiving on disks, it's a good idea to also maintain backup tapes of that archive, he said. That way, your client is covered in case one medium ends up being unexpectedly vulnerable in a certain way -- for instance, susceptible to a new kind of attack or found not to be fully WORM. The exact value in this duplicate approach is hard to express, but many companies like the ease of mind it provides, Crump said.
Your client should also consider a tiered, multi-target system in deploying an email archiving system, Benton said. For instance, a company may archive messages after one week to a fast, on-site drive, and then move them to second-tier storage after a month and finally off-site after a year. It's important to make sure your client integrates its archiving policies with its overall data lifecycle management approach, Benton said. On the other hand, make sure that old archives aren't being constantly included in your client's regular backups; email archives from a year ago don't change, so there's no reason to keep transferring them to the backup every week.
Because regulatory compliance is one of the major drivers for archiving, a big part of deploying an email archiving system is making sure that it fits the company's legal needs. As with any policy-driven technology, it's also important to make sure it works with your client's corporate culture. In the final installment of this Hot Spot Tutorial on email archiving, we'll show you best practices for working with your client to ensure the email archiving system achieves its goals.
This was first published in October 2008