Thanks to markthshark for his excellent diary today regarding the White House email system and their claims that they couldn't restore the data.
Fortunately, I happen to have a little technical expertise on this matter. I spent about 5 years doing support, administration, and design for one of the biggest Lotus Domino installations in the world. (Domino is the name of the other "half" of Notes. It runs on the servers, and Notes is the client that runs on users' workstations.)
The installation is part of the banking industry, which thanks to regulations like Sarbanes/Oxley and typical litigation support and internal investagatory work, has to regularly do retrieval and searching of archived data from previous years.
In my expert opinion, the White House is misrepresenting almost everything about this situation.
Let us decend softly into geekery. First, on the "archaic" software in question:
simply a computer glitch that ensued when the White House wanted to phase out an archaic email program.
In an exchange with Payton, Issa characterized Lotus notes as "wagon-wheel" technology.
"I wouldn't want to do business with somebody still using Lotus Notes or still using wooden wagon wheels," Issa responded.
Issa person clearly has no clue what he's talking about. I'm no apologist for IBM in this matter -- Lotus Notes isn't the best or easiest-to-use email client around. But Lotus Notes isn't an email program. It's a database, and its primary built-in database application happens to be mail. Lotus Notes has a lot of weird things about it for an email program but which make a lot more sense if you view it as a database and document workflow system.
Companies that use Notes/Domino for the strengths it has and for which it was designed -- the ability to build applications that tie into other applications and the email system seamlessly and allow for centralized access control with the same system -- tend to quite like it. Those that buy it expecting email don't. And I don't blame them one bit, but it's more their fault than the software's.
That said, Notes is by no means archaic. IBM has maintained a robust release cycle which no large organization can even pretend to keep completely current with. The family of related software has also grown apace. Congressman Issa is either ignorant of the facts, or has chosen to ignore them for political gain.
On to the subject really at hand, now, the recovery of data. I'll put this as simply as I can:
As long as there is full backup of the system on which the Domino servers existed, which is the case for any tape archival scenario, there is no reason that the data cannot be recovered.
I've done this, I'll even explain how, because in high-level steps, it's easy to follow:
- You either restore the Domino software from backup along with the Domino server data; OR you install Domino afresh on a server, and with just a small ID file or two, you give it the identity of the server it was in the past. I've done it both ways.
- You stop the server (Domino) software from running
- You restore all the database (mailbox) files
- You start the server software
Hey, there you go. Extracting the data from the files is at that point a fairly simple exercise, depending on if you want to filter it by some criteria or other.
Now, granted, this is just a SNAPSHOT of a point in time. I will allow that getting a record of all messages in and out that were not present at the time of system backups could be either prohibitively difficult or impossible, depending on the architecture of the Domino system and the backup and archival solution. But to claim that NOTHING exists is not only wrong, it's laughable.
I'll also allow that the cost of recovery COULD be higher than markthshark allows for -- not due to Notes/Domino -- but depending on the archival system. Reading between the lines, I'm guessing their system was on an AS/400 or other large system, which may have been using an archival system that is only possible to recover back to an AS/400. I'm less intimately familiar with this side of the operation, so there's some speculation in this.
Even so, these systems can be and are often leased or rented, and the mailbox files, once extracted back to some sort of disk, could be transferred onto a medium readable by an inexpensive UNIX or intel-based system. The $500,000 purchase price is irrelevant.
Some large UNIX systems can cost that amount, or even higher, but again, this is a smokescreen -- systems running Domino are not that powerful to serve the internal needs of shuffling email, but the requests from users -- expected user count is the primary factor in determining the sizing of a server system. In a recovery-only scenario, the user count is essentially nil, so a minimal system could carry it out as well as a huge system, if more slowly.
I'll finish up with a touch on the hard drive issue, with regard to Notes. Notes itself allows you to make a local copy of your mail database on your local machine, but does not require it, and in fact, the system settings allow the administrators to disable this ability. However, this is always a secondary copy -- regardless of whether or not the hard drives have any useful data in this respect, there should certainly be a tape archive that contains it.