Answer honestly. How
often do you use search engines? I
know. Thats an incredibly broad
question. Sure, I use Google, MSN or Yahoo
to find information on web sites just like millions of other web surfers. But how often do you use Enterprise Search tools to find data relevant to performing your
job? That is, disparate data stored not
only on the Internet and your intranet, but across data and email repositories,
content management systems and databases to name a few. Admittedly, the closest Ive come to using an
enterprise search tool is Google Desktop or our companys intranet. Finding files on my workstation or links on
the Internet? Yawn. I want to see something that helps me findthat long lost box of CDs in my garage.
Okay, there is nothing that can help me find what Im
looking for in my garage. (In fact, I
think my wife secretly threw the box away and gets great enjoyment from
watching me fruitlessly search for it.) So,
short of that miracle, there are some very intriguing advancements in search
technology that will become more important as the volumes of data whichcompanies store continues to grow.
When you stop and consider all the places your company
stores data, its daunting to think about how long it would take to search through
each individual location. Todays
enterprise search tools can not only search every nook and cranny of your
network, but can search for data regardless of the format its stored. Tools such as Vivisimos Search Engine can search for
data within relational databases, PDF, ZIP and GZIP files, email, MS Word,
Excel, file servers, intranets. The list
goes on and on. Search results arent
returned in the common linear lists which make it difficult and time consuming
to sift through either. Instead, premium
search engines can offer results in clustered and conceptual categories. Take a look at some of the online demos by Vivisimo and you
immediately get that Aha! reaction about what should be the norm when youexecute a search.
Another major player, if not THE player in enterprise
search, is a company called Autonomy. It incorporates what it calls Meaning Based
Computing which takes unstructured and disparate data, and enables computers to
create meaningful relationships. This
goes beyond just performing a keyword lookup which only works to find all
documents with an exact match. Instead,
Autonomy also attempts to find results which contain the same or similar
meaning to what is queried and can execute searches based on pattern matchingand language translation.
Many corporate solutions dont rely on only one search
technology. This is where federated
search systems come in. Federated search
takes one familiar front-end search box and passes the query to multiple search
engines. So, say, a search on your companys
intranet may use Vivisimo or Autonomy technologies to locate data within your
network, while a hook on the backend may also return results using Google or
MSN. This could also have applications for
companies who wish to point search engines toward data repositories on company
extranets. For instance, a search for a company
purchase order would not only return results from your corporate CRM system,but also the invoice stored on the outside vendors data storage system.
Yes, there are also enterprise search engines such as Fast Search & Transfer (FAST) which
cater solutions toward specific niche markets such as corporate regulatory
compliance, criminal investigation and litigation compliance protection. FAST search engines even have the ability to
search video and audio formats and return results based on closed caption text,voice-to-text recognition and sound recognition.
Speaking of niche search engines, Splunk is aimed at IT analysts who have spent
countless hours searching for the source of an IT problem. Pretty much nailed all of you, didnt
it? Splunk allows you to search all data
being logged by any service, application or device within your data center,
regardless of source or format in real time.
When an IT problem is reported, Splunk aims to reduce the mean time to
recovery through live reporting. Perform
a search for a specific error message or incident type and turn it into analert that constantly monitors for the event and notifies you when it occurs.
I am pleasantly surprised at the advancements being made in enterprise
search technologies. With the sheer
volume of data being collected and held onto by companies, it makes sense that
IT has rushed in to fill yet another void.
And since the main hurdle of how to index disparate data types seems to
have been conquered, the direction now seems to be focused on what to do with
the search results returned. This is why
so many of the companies mentioned also offer collaboration with Business
Intelligence tools to provide greater data analysis. This should be interesting to watch as the
market continues to mature. Now, ifsomeone could please help me with my garage