Data Warehousing update – part 2 – DATAllegro

This was going to be an update on recent product announcements
from both DATAllegro and Netezza but I have now concluded that
there is too much discuss to cover all of this in just one
additional article so there are going to be several. I don’t know
just how many. And I am planning to add updates for Vertica,
Greenplum, H-P and Dataupia; and with Teradata’s user conference
upcoming (it may even have happened by the time you read this)
there may be even more than that. It may even become a rolling
series of updates. Anyway, this one’s about DATAllegro.

DATAllegro made two significant announcements in August, one
releasing what it called (at that time) an Archive Appliance and
the other introducing support for grid-enabled appliances.

As far as the Archive Appliance is concerned this is basically
identical to DATAllegro’s standard offering except that you use
higher capacity disks, so that you can have up to 200Tb in a rack
(and you can have multiple racks) with fewer CPU resources (one per
node instead of two). You can also encrypt the data if you want
to.

I think this is excellent. But it’s not really an archive
appliance (which is probably why DATAllegro has stopped using the
term) it’s more of an anti-archive appliance. Anybody who read my
recent article on the value of data will know that in many
instances it is valuable to maintain data on-line for analysis
purposes but you often can’t afford to do that, so you push it off
on to tape, which is what most people think of when they hear the
word “archive”. What DATAllegro is doing here is allowing you to
keep this data on-line so that you can analyse it at will, albeit
with slightly less good performance than you would get with a
standard DATAllegro appliance. As far as I know DATAllegro is the
only data warehouse company doing this at volumes in excess of
100Tb and the only company doing it, in particular, in the Teradata
space.

As far as the grid support is concerned, the idea here is that
you can have multiple appliances integrated with one another in a
distributed architecture based on a hub-and-spoke approach. Note
that this might include both regular and (anti-)archive appliances
and can also be used to provide disaster recovery from a remote
location. Where multiple active systems are in place, DATAllegro’s
Intelligent Query Router (which used to be an optional extra but
which is now bundled into the system) is used to distribute queries
across appliances in an optimised fashion.

More recently, DATAllegro has made a third major announcement,
which is that it is partnering with Bull in Europe. Specifically,
Bull will be launching its Data Warehouse Parallel Server (DWPS),
which is basically the same as DATAllegro’s own product with the
exception that instead of using Dell servers throughout the
appliance it will instead be using Bull’s NovaScale Servers. Bull
will then be marketing and selling the DWPS throughout continental
Europe.

This is potentially of great importance to DATAllegro in that,
at a stroke, it adds a significant sales force. Moreover, this
sales force already has strong customer ties and can be expected to
introduce DATAllegro to a much wider audience than was previously
possible.