Informatica unveiled their newest product in the PowerCenter line, the PowerCenter Express, at Informatica World this year (find Smartbridge’s experience of the convention here).
The
sales pitch is certainly catchy: Free PowerCenter! First I heard of it,
I wasn’t sure what to think – is this a marketing gimmick? What’s the
catch? But hey, at that price, it is easy enough to find out by one’s
self, and that is precisely what I did. And color me pleasantly
surprised.
The Limitations of PowerCenter Express
To
their credit, Informatica is upfront with the limitations of the
product. This is a good thing – no easier way to shoot yourself in the
foot than sneak small print on your clients under a guise of a
no-strings-attached free download.
As you could expect,
PowerCenter Express is not PowerCenter – the free Express version can
only process a quarter of a million rows per day – good enough for small
development, but it is best considered a demo version. The paid version
includes multi-user support, and removes the processing limitation, but
is still limited to five users and no job parallelization.
If
your company is already using PowerCenter, you are probably long past
the point where you could realistically choose to downsize to Express.
But if your company was too small for the behemoth that is PowerCenter,
then Express may be exactly what you need.
I suspect that
Informatica sees Express as a way to reach to clients that, until now,
were too small to warrant their larger products – maybe a way to get
them to dip their toe in the waters.
Do
not think, however, that this is “PowerCenter lite”. Express is a
product on its own right (the paid version, more so than the free). A
small-to-medium company that finds itself in need of an ETL can do much
worse than invest in Express. Even when I was building PowerCenter ETLs
for a large bank, we seldom ran more than two or three medium jobs in
parallel – the strain it puts on the source and target is just not worth
the time savings – and the larger jobs usually ran on their own.
The
lack of parallelization will hit only if you had a large number of
small jobs; and even then, serializing them shouldn’t be more than a
small inconvenience, although not having faced the actual issue, do take
this prediction with a grain of salt.
Express Installation
Grabbing
a copy and installing it was simplicity itself. I have always felt that
PowerCenter’s greatest strength is its ease of use, beyond even its
connectivity. I’m happy to see Informatica expand the ease of use to the
installation.
A stand-alone install program is all it takes to be
up and running. I was building my first test mapping less than an hour
after deciding to download Express, and ran it successfully in less than
two hours (it wasn’t a very interesting mapping, admittedly, but it was
a reasonably complex join of flat file data against a local database,
aggregated and sent to a remote location – the kind of “simple” ETL that
has been known to cause me headaches when attempted in unvarnished
SQL).
One word of caution: Express is not a toy.
Even the free version has a fully functional PowerCenter server. When
turned on, my laptop went into permanent spin, and my memory and CPU use
climbed several notches. I found myself turning it off just to give my
poor laptop a break. It worked for testing, but if you are going to use
it to develop an actual ETL, consider installing the server portion on
an actual server.
PowerCenter vs. PowerCenter Express
PowerCenter Express is by no means ‘lite’.
As
a long time user of PowerCenter, this part is actually tricky to write.
How many of the changes are “bad” and how many of them is just me being
an old curmudgeon? It’s difficult to say. The good news: you needn’t
worry. They did not strip PowerCenter down. Every transformation you can
find in “classic” PowerCenter is in Express as well.
Express even
includes a bunch of direct connections to social media to speed up your
mapping development: Twitter, LinkedIn, Facebook, you name it. And I
loved that they finally dropped the “source” and “target” – an
unnecessary distinction, when most external entities end up being both.
Express automatically assumes that, and the whole is more compact for
it.
I am less happy about the lack of Sessions. They are not gone
completely – the workflow is still a sequence of objects that are
associated to mappings – but without my usual central point for
redefining sources and targets, I was left scrambling to find where to
do so. I suspect this is more my muscle memory that led me to looking in
all the wrong places, though. As always, F1 brought up the help, and
once I had read the manual, it became easy again.
There are a few
other nits I could pick – I am not entirely convinced I like the new
graphics, the ribbon or the “all in one” approach – and I cannot even
guess at what other differences I would eventually find, if given enough
time, but these are minor.
Express is PowerCenter, and the old
approaches to mapping design will still work. It is still visual,
intuitive, and easy to use.
So Does Express Pass the Test?
If
Express’ name wasn’t attached to Informatica PowerCenter, I’d
considered it a basic ETL, with potential for growth and useful mostly
for small deployments.
The equation changes, though, when you
consider that if you do outgrow the capabilities of Express, you can
easily upgrade to PowerCenter. It is an interesting approach, and I
could almost say Informatica has managed to square the circle.
This
first visit to the tool has proven successful enough that, were I to be
required to use Express as the ETL tool, nary a complaint would escape
my lips – and those of you that have met me know how rare an occasion
that is.
Disclaimer: The company I work for is partners with several vendors including Informatica
Disclosure: I am a real user, and this review is based on my own experience and opinions.
GaryM, I understand that automatic refresh can cause data mapping errors, but identifying the changes, notifying the users and letting the user decide how to apply the changes (update, ignore, create a copy of mapping and edit ...). When the object in DB changes, chances that mapping will fail are high depending on the type of change, why not know about it and prepare for it before hand. Either you need to run a SQL and identify DB changes periodically and apply them in Informatica manually or Informatica identifies them and let's users decide (when repository reconnects).