Two novel ideas for keeping mail storage to a minimum

[furball4]

I'm sure someone has thought of these before, but I've never heard of them being implemented.

The Problem
Email that was only relevant the day it was sent and has *zero* long-term value is nevertheless stored for years, backed up a zillion times, lovingly restored alongside today's mail after a loss, etc.

Solution #1
Allow and/or force users to choose a length of time that a message will remain relevant when it first comes into their inbox. For a meeting notice they can hit 'One Week'; for an order confirmation 'One Month'; for an HR form 'Three Years' or 'Forever'. Then they can forget about less effective methods of cleanup, because the mail will be discarded after its period of relevance has passed. They can do the same for sent messages, to keep their sent folder respectable. And, in an integrated client/MTA system like Zimbra, that setting could even be used as a preset for recipients in a custom header. So when someone (or their calendar) sends out an email about an event that is going to happen on Tuesday, the recipients see it marked 'Relevant for one week only' by default.

Solution #2
Each day, users are prompted to review their correspondence from 90 days ago and separate the keepers from the trash. This is low-flow, has the hidden advantage of reminding people about things they may have forgotten, and seems like it could make a huge difference in the amount of irrelevant email being stored, either alone or in combination with the first solution.

[cdamon]

In general, this isn't really a problem because storage is cheap and plentiful. Knowing that, many users prefer to keep everything around, just in case. It's more of an issue in terms of managing your mailbox, and keeping what matters to a user in front of them.

I've heard a slightly different take on #1 before, where a user could mark a message/conv for deletion after a certain amount of time (day/week/month). Then it automatically gets moved to Trash after that time. That's helpful for users who are running up against a quota, or who keep everything in their Inbox.

#2 isn't feasible, in my opinion. I would want to strangle a system that makes me deal with old mail every day.

-Conrad

[fmodola]

do you think the first solution proposed can be integrated in next releases of ZCS ?

I think it would be a good solution to automatically mark mails via the filter rules, so that users can tell ZCS to mark for deletion mails by types of mail (new appointment notification, mails from a specific service or with a specific subject ... etc), specifying the time before deleting them.

[furball4]

In general, this isn't really a problem because storage is cheap and plentiful.

I understand that sentiment, but I disagree. Hard drives are cheap and plentiful. But storage is still expensive and complicated from a TCO perspective. You have to have a machine to put it in, power it, cool it, maintain it... but most importantly you have to back it up. If you have a solid disaster recovery plan then you are moving copies of it offsite as often as daily or even instantly, you are doing regular recoveries for testing purposes, and recovery time is one of the important metrics. When all of that machinery is added into the term 'storage', the cost of the hard drive all but disappears. But the difference between, say, 8GB and 80GB of email remains huge.

[KevinH]

Originally Posted by furball4

But the difference between, say, 8GB and 80GB of email remains huge.

I really don't think 8GB and 80GB are a big deal or even 800GB. Disks these days are easy to come by in the 100's of GB these days. Zimbra has the ability to keep the new/recent mail on expensive/fast disks and then move older email off to cheaper storage. This feature HSM makes it super simple to offer user's multi-GB mailbox quotas and no need to worry about storage costs. If your biggest problem these days is storage costs you should look again. Things like end-user help desk support make disk space costs a non-issue. I think your ideas have merit but to be honest there are higher priority items in the queue. Please file these in bugzilla so folks can vote on them and get them into a future release.

[furball4]

Well again, I'm not talking about storage in one place only. I'm talking about storage that is part of a strong DRP. If that includes offsite storage, which it does for many companies, then the difference between a few gigs and a Terabyte is massive. Even if you can do incremental backups, you need a restore all at once. Maybe it wouldn't grate on anyone else that 70% of that restore time was spent restoring useless, outdated communications, but it would grate on me! I don't want to store trash if I can help it. I don't want to delete most personal emails, ever. So that is a greater reason not to keep trash. If I'm not going to sunset email at the 5 or even 10 year mark, then I had better keep things well-pruned. I don't know... maybe people have just given up on that concept at this point.

I agree that there are much better things to be adding right now, though!

[Ericx]

We are trying to have a rock solid disaster recovery plan. Any company that thinks they have one, either does not have much data or is just kidding themselves.

Although hats off to many of the business in the twin towers after 911. The business I read about were back up (from an IT standpoint) within 12 hours. But that is the exception not the rule.

My problem is not email but attachments! I could back up all our "email" on a thumb drive. Our company deals with Autocad drawings. 10 Megs here and 15 megs there... it all adds up real quick. I have a friend that works at the City and their new compliance rules (from FEMA or someone) require them to backup everything including spam. Over 70% of their traffic is spam! What a freaking nightmare. I'm glad my tax dollars are going to backup spam!

I've been thinking out a way of having the server strip out attachments and send the end users links of where the files are attached. I may never get around to it, but I'm thinking about it.

[fmodola]

I don't think a SAN storage is cheap, and I do agree that we only need to keep really important emails and documents.

Particularly, when you're invited, and appointment is sent, but after the meeting, there's no need to keep the appointment emails ... there's already an entry in the calendar.

When you send an email, it could be for an urgent problem, so after the emergency, I don't think the email need to be conserved.

When you receive this type of email, I could be great to do the same operation.

[chh]

What you have to consider is that, in a lot of countries, you have to keep EVERY email that has bussiness value. And you may not allow the user to decide if a mail has to be kept. So basically you have to keep every mail if you want or not.

Besides that: The question is not if you should keep the mail but how.
The only feasable solution in my opinion is archiving "old" mail.
Most useres only seldom touch mails older than 2 or 3 weeks. If these are moved to cheaper storage (like Zimbra can do) You can decide how often to back this up. Or if you use "real" archiving features (hope those are implemented quite soon in Zimbra) you could even use WORM file systems that you need to backup only once.

[dkarp]

Originally Posted by chh

Or if you use "real" archiving features (hope those are implemented quite soon in Zimbra) you could even use WORM file systems that you need to backup only once.

Does the Mail Archiving feature in Zimbra 4.5 Network do what you need? If not, what are the unfilled gaps between that feature and what you require?