Information Technology Reference
In-Depth Information
4
Scheduling Methodology in DMFB
The basic microfluidic operations namely merging, mixing, splitting and storage is so
far accomplished on predefined fixed module locations comprised of a group of cells
within the 2D planar array. In order to avoid conflicts between droplet routes and
assay operations, a segregation region is defined around the functional region of each
such microfluidic modules. Such locations dedicated for microfluidic operations are
considered to be non reconfigurable and fixed. These cells together with a segregation
zone consumes considerable area within the DMFB - leaving the overall area for
reconfigurable operations e.g routing to be confined within narrow channels. For
example mixing is performed by bringing two droplets to the same location and
merging them, followed by the transport of the resulted droplet over a series of
electrodes. By droplet movement, external energy is introduced, creating complex
flow patterns (due to the formation of multilaminates), thereby leading to a faster
mixing [26]. Mixing through diffusion, where the resulting droplet remains on the
same electrode, is very slow. Hence we can conclude that it may be possible to
execute the mixing operation at any available area comprising of the 2D array
configuration of electrodes necessary for mixing. The operation of dilution carried out
through a sequence of mixing and splitting may also be executed dynamically in a
similar manner within the 2D array. Thus most of the basic microfluidic operations
can be rendered reconfigurable and executed dynamically as per schedule and
placement plan through routing -where one to three electrodes are involved at any
given time. However there are certain “non-reconfigurable” operations required to be
executed on real devices, such as reservoirs or optical detectors which calls for
reservation and binding of fixed resources in the form of individual or segregated
cluster of cells.
The conventional architecture level synthesis starts with a sequencing graph
representing different assay operations with their mutual dependencies. Each node in
the sequencing graph represents an operation and each directed edge represents
operation precedence and sequence of execution of each operation. Next, scheduling
and binding assigns time-multiplexed steps to these assay operations and bind them to
a given number of devices so as to maximize parallelism [27]. In scheduling and
binding, each operation will have a set of devices in the form of module library being
available for resource binding. Table 1 presents the results of the experiments
performed in [26], where several mixing times were obtained for various areas,
creating a module library. Figure. 4, shows the type of mixing and the array
configuration for modules available to choose. Choice of different modules may result
in various reaction area and execution time. Based on the execution time for each
module, the start and completion time of each operation are arranged. On the basis of
the scheduling result, device placement and droplet routing are conducted to generate
a chip layout and establish droplet routing connections between devices in a
reconfigurable manner [28][29].
