Google’s Mini Search Appliance may be too Mini
Google have started selling their mini search appliance. Much like it’s larger gold brother, this is a LAN file indexing machine.
What is concerning however are the file limits that google imposes. Currently the mini appliance supports 50, 000 documents. Many of you who have been using computers casually may not see the relative “lowness” in this value. On the other hand, the likes of Prejudice and I have real trouble coming to terms with this limitation.
You can go years without producing this many files it’s true. In fact recently I lost 46GB of files due to a hard disk failure and this certainly set a few things back. C’est la vie, there is little we can do to protect ourselves from failure except back things up. Well when the backup goes down at the same time as a main fault, life just gets harder.
Still, most of my files were backed up safely and so I didn’t loose too much that was of major importance. Despite my loss though, these google appliances are not enough for my desktop let alone an entire LAN. In fact, scrap the desktop idea entirely, it doesn’t stand a chance. How about my documents folder? No, it can’t even fit mine in there..
If you don’t believe me, try looking at this screen grab:
Now clearly this is over the limit of the mini appliance. After looking at virus scan results (which yes, index more than documents, but system files are only so many) I can safely say that I can encroach upon 1/15th of the Google Search Appliance. Between Prejudice and I we can fill 1/7th of it. So what, you can have 15 high end users who carry running archives? :-/ This value seems somewhat low. Now of course, we are the types to keep references to all kinds of stuffs. We are also programmers with large resources of source files and utility files we have written. Still… 15 people.
I’d love to see one of these things in action. I’d also be interested to know (in real terms) how much faster it is than running say the MS Indexing service with a distributed capture. The file type support is useful, although I am unsure of how many plug-ins are available for the indexing service but I am sure it would be possible for many many formats to be added in a relatively short space of time given a bit of development attention.
If you look at the success stories for the devices you generally find that they are designed to tackle not a whole intranet as such, but normally specific areas. Sales being one such area that had more than one instance. I wonder if we could justify the cost purely to reduce the need for organisation and replace it with indexed search? At that price, maybe next year. If the mini was capable of a few million files then it would certainly gain far more consideration.