Database Reference
In-Depth Information
IT for IT:
Data Center, Machine Data, and Log Analytics
Hadoop-based log analytics has become a common use case, but that doesn't
mean it's deployment is as widespread as it should be. Log analytics is actually
a pattern that IBM established after working with a number of companies,
initially in FSS. We've since seen this use case come up across all industries; for
that reason, we'll call this pattern
IT for IT
.
Big Data-enriched IT for IT helps clients to gain better insight into how
their systems are running, and when and how things break down. For example,
one financial firm affectionately refers to the traditional way of figuring out
how an application went sideways as “Whack-A-Mole.” When things go
wrong in their heavily SOA-based environment, it's hard to determine what
happened, because more than 20 systems are involved in the processing of a
given transaction. (We've all seen this movie, everyone running around the
war room saying, “I didn't do it!” There's also a scene in that movie where
everyone is pointing fingers…at you!)
One client we helped with this usage pattern ended up with the capability
to analyze approximately 1TB of log data each day, with less than 5 minutes
latency (this use case is just as applicable with much larger or smaller log
generation rates). Today, this client is able to decipher exactly what's happen-
ing across their entire IT stack, within each and every transaction. When one
of their customer's transactions, spawned from their mobile or Internet
banking sites, fails, they're able to tell exactly where it happened and what
component contributed to the problem. As you can imagine, this flattens
time to resolution metrics.
This client could take the insights that they've harvested at-rest, and leverage
Streams to investigate the health of their networks in real time. For example, if
constant consumption of a memory heap in the Hibernate layer is strongly
correlated with a stack overflow in the application server, being able to pick
up this issue at its genesis could save the transaction or prevent the network
from experiencing an outage. As an example, one of our telecommunications
customers uses Streams to analyze machine data in real time to find misbe-
having mobile computing applications that are harming the networks, so the
offending applications can be terminated.
