Friday, February 24, 2006
Gmail for domains
An interesting post mentions that Google is now experimenting with gmail for other domains. I expected them to market their gmail product as a service, and I believe this would be very valuable for many companies to do their (internal) e-mail with gmail.
In addition to the added value for the users, I am convinced that the approach of gmail not to delete mail will eventually use less space than conventional e-mail systems. And this means reduced cost for storage, back-up and systems management (by users and system administrators). Saving space by keeping everything is not intuitive, so that needs some explanation.
With conventional e-mail systems, the sender of a reply knows that often the recipient will not be able to recall the original mail when reading the reply (because it was not kept, or because it is hard to find). The convention thus is to reply "with history" quoting the original mail in full. This is convenient for both the sender of the original mail (to understand the context) and for the sender of the reply (to archive mail and reply together). Now in theory both participants could go and discard their copy of the original note, if needed.
Now imagine a conversation which takes a few more participants, where different people reply in different stage of the thread causing the discussion to fork at different points in time. It is not uncommon for most participants to end up with a pile of notes, each not containing a subset of the mails exchanged. It is almost impossible to reduce this to a single copy of the entire thread. So when the thread consists of n mails, each participant ends up with n*n mails archived. And since often the number of pariticpants in a thread grows with the number of mails, the total cost for a company to archive becomes n*n*n.
Clearly, when the size of your mail archive grows that fast you will have to start removing mail. This will encourage people even more to reply "with history" and increase the size of your archive, and so on.
A company with thousands of employees has few options to address these problems but limit the storage on central servers and have the users archive the remainder on their local workstation. These archives may not be backed up, may get lost or end up in wrong hands when the laptop is stolen.
I find it amazing how something "simple" as just archiving all mail helps to solve this problem.
An interesting post mentions that Google is now experimenting with gmail for other domains. I expected them to market their gmail product as a service, and I believe this would be very valuable for many companies to do their (internal) e-mail with gmail.
In addition to the added value for the users, I am convinced that the approach of gmail not to delete mail will eventually use less space than conventional e-mail systems. And this means reduced cost for storage, back-up and systems management (by users and system administrators). Saving space by keeping everything is not intuitive, so that needs some explanation.
With conventional e-mail systems, the sender of a reply knows that often the recipient will not be able to recall the original mail when reading the reply (because it was not kept, or because it is hard to find). The convention thus is to reply "with history" quoting the original mail in full. This is convenient for both the sender of the original mail (to understand the context) and for the sender of the reply (to archive mail and reply together). Now in theory both participants could go and discard their copy of the original note, if needed.
Now imagine a conversation which takes a few more participants, where different people reply in different stage of the thread causing the discussion to fork at different points in time. It is not uncommon for most participants to end up with a pile of notes, each not containing a subset of the mails exchanged. It is almost impossible to reduce this to a single copy of the entire thread. So when the thread consists of n mails, each participant ends up with n*n mails archived. And since often the number of pariticpants in a thread grows with the number of mails, the total cost for a company to archive becomes n*n*n.
Clearly, when the size of your mail archive grows that fast you will have to start removing mail. This will encourage people even more to reply "with history" and increase the size of your archive, and so on.
A company with thousands of employees has few options to address these problems but limit the storage on central servers and have the users archive the remainder on their local workstation. These archives may not be backed up, may get lost or end up in wrong hands when the laptop is stolen.
I find it amazing how something "simple" as just archiving all mail helps to solve this problem.
