| Craig S. Mullins
Pervasive Computing and Pocket Databases
entering your readings into this gadget, here,” replied the
technician. “We’ve had to do this for the past few weeks or so.
It’s all about some new procedure for storing patient’s vital
isn’t it only useful for you?” Scott asked.
“I mean, if you’re just entering it into your gadget, then
no one else can use the data, can they?”
it doesn’t work like that. When I’m through entering the
information – like blood pressure, temperature, height, and weight
– I can send it to the doctor’s gadget, too. All I have to do is
just point it at the doctor’s gadget and select this option to
“beam” the data to the doctor. I carry this thing with me all day,
recording patient information. But before I go home for the day I have
to put this gadget in a little gizmo and press a button. The next day
doctors, nurses, and I can pull your information up on our central
cool to me,” Scott said, translating the word gadget to PDA and
gizmo to docking station in his mind. “Do you have any idea how it
But it makes life easier for me. I don’t have to worry about using
that infernal PC system we used to use. It was always messing
something up and we’d have to write everything down first, and then
enter it into the computer. That wasn’t smart. And now we don’t
have to worry about reading the doctor’s poor handwriting or losing
file folders any more. I love it.”
was intrigued. “Does everyone use them?”
we all use them now. Even the nursing staff uses them when they make
home visits to the elderly and disabled. And we have outfitted the
emergency crew in our ambulances with the devices too.”
is a practical application where an electronic gadget is making life
easier,” thought Scott. “I’ll have to remember to tell this to
my wife when she complains about that Palm Pilot I just bought.”
of us in the database management profession spend our days dealing
with rapidly growing enterprise databases. These databases are used to
support OLTP systems and e-business applications. As the usage of
these systems grows, inevitably the data in those databases grows. But
the example above outlines another growing trend – smaller,
individual databases stored on PDA devices like Palm Pilots and
will probably never completely abandon the ever-growing, ever-more-
complicated large enterprise databases upon which most of our
businesses are based. But we will increasingly be supporting smaller,
mobile pieces of those databases. Our culture and demand for immediacy
will necessitate this trend. As our reliance on data and technology
increases, our need for disconnected (or intermittently connected)
devices will increase. Users of the technology simply will not
tolerate having to be plugged in all the time.
call this pervasive computing. The overriding feature of this type of
system is its portability. The devices can be carried wherever the
user has a need to use it. And most of these applications will require
persistent storage of data. In other words, they will need a DBMS.
Database Management Systems
of the major DBMS vendors supply versions of their products that run
on the popular PDA devices of today: the Palm Pilot and the Microsoft
Pocket PC. IBM’s DB2 Everyplace, Oracle8i Lite, and Sybase Adaptive
Server Anywhere are prime examples.
we may be stoking the fires to keep our large enterprise databases
operating efficiently and providing service to our data warehouses,
OLTP systems and e-business applications. But we are also going to be
slicing those databases up into smaller, subject-oriented, personal
pieces for deployment on pervasive computing devices. And we will need
to manage the flow of data into and out of these applications.
biggest impact will be caused by trying to manage data synchronization
from numerous individual PDAs to the centralized database. When should
synchronization be scheduled? How will it impact applications that use
large production databases that are involved in the synchronization?
How can you ensure that a mobile user will synchronize his data
reliably and on schedule? If remote changes are not synchronized how
do you deal with centralized databases that do not contain up-to-date
data? These are the questions to be faced by DBAs encountering
pervasive computing applications.
you thought you had your data flows all under control, didn’t you?