wissel.net

Usability - Productivity - Business - The web - Singapore & Twins

Cost of messaging storage not an issue?

David Ferris has a blog entry titled " Cost of Exchange Storage Not An Issue". What he says is not Exchange specific (it is only the first time that Exchange would be capable of supporting large mailboxes) and can be applied to any messaging system. David states:" Users will have large mailboxes. 5GB to 20GB will be common". With storage cost (again David's figures) of $2-$30 per 30 GB that translates in per user/per year cost in the single digits. I'm not so sure about these figures. They might be true for home storage, but for large scale enterprise storage systems David might be off by more than one dimension. Storage cost isn't a linear figure, but has increasing marginal cost once you reach size limits.
What's your take. Are large enterprise mailboxes coming?

Posted by Stephan H Wissel on 08 December 2009 | Comments (5) | categories: Software

Comments

posted by Kevin Pettitt on Tuesday 08 December 2009 AD:
Users are lazy, and will find the path of least resistance to getting their work done. You can use carrots and/or sticks to break them of over-reliance on email. You can put technical band-aids in place to mitigate the infrastructure impacts of such over-reliance (e.g. DAOS).

The "stick" approaches can help (e.g. quotas), but this whole endeavor will be an uphill battle until an effective carrot is found that makes users *want* to avoid email because it is no longer the easiest way to get work done. Facebook's success may offer lessons here. Maybe Lotus Connections will play a big role in achieving similar results in the enterprise. Quickr I view as more of a stick because, while putting files into Quickr might make life easier *tomorrow*, it is still slightly harder *today* so laziness kicks in.

So, unless a technological paradigm shift occurs (soon) in the enterprise that matches Facebook's impact on personal communication patterns, gigantic mailboxes are probably inevitable.
posted by Keith Brooks on Tuesday 08 December 2009 AD:
yes and no. Many companies don't want users to have out of control massive mail boxes because of many reasons.
DAOS of course helps part of this.
The other problem is server performance, drag time, inbox lookups (many multiple GB people have 1,000's of emails in their inbox). Not to mention time of backups, running server tasks becomes a long event too.
So yes we all may get bigger mail files but it may not be better.
posted by Philip Storry on Wednesday 09 December 2009 AD:
Anyone who thinks storage is cheap isn't buying much of it.

By which I mean they're looking at the price of individual disks - not arrays, and much less SANs.

Storage costs are certainly going down, but they're not "cheap". The hardware, the labour, the time out of production... It all adds up.

And woe betide you if you run out of physical room in your SAN. That can be very expensive.

Storage is cheaper than ever, yes. But it ain't cheap yet, and it's a long way from being so.
posted by Rupert Clayton on Wednesday 09 December 2009 AD:
There's a valuable conversation to be had here, but David Ferris doesn't really seem to be driving it.

Alan Elliot's comment { Link } on David's posting lays out many of the reasons why David's approach is far too simplistic.

I do think that DAOS will reduce the rate of storage growth for Domino shops. And DAOS is a much better fit for moderate performance RAID 5 arrays than Domino databases.

Also, tiered-storage archiving approaches will help some enterprises keep a further lid on storage. Although regulatory and e-discovery issues are likely to be as much of a driver there as storage management.

I do see a welcome trend toward IT departments managing mail storage centrally, rather than forcing users to manage this through aggressive quotas and client-side archiving.

For Domino, one of the areas where contention will likely remain is the size of the Inbox. Admins are eager to reduce the indexing load by using the Inbox management agent in Domino 8.x to silently drop old messages from the Inbox. Users won't like this -- they don't expect stuff to disappear without notification -- so it will need to be carefully sold to them.

Although I can see the merits of both shared content management (e.g. Quickr) and corporate social networking (e.g. Connections), I just can't see that having a big impact on e-mail traffic in many companies any time soon.

All of this means that the dire warnings about Exchange 2010's storage requirements should be given some credence. While Domino is delivering technologies to restrain mail file growth, Exchange 2010 does seem likely to speed that growth.
posted by Henning Heinz on Wednesday 09 December 2009 AD:
Alan's comment is interesting although I am sure some CTO will ask why companies like Google charge 50$/year for tons of gigabytes and internal storage cost 750$/user per year?
Then you might talk about security, reliability and overall performance. That might work and has some truth. The problems appear when for example governments start moving their data to the cloud (or get the go to do so). So maybe David Ferris should have said that internal messaging systems must support large mail files at a reasonable cost to stay competitive. Also with all the regulations nowadays, quotas and small mail files are often not saving big money because you have to store mails for a long period of time anyway.