Biomedical Engineering Reference
In-Depth Information
Implementation
The National Science Foundation (NSF), Department of Energy, and other government agencies,
often in collaboration with industry and academia, have virtually unlimited resources available for
developing and maintaining networks that have bioinformatics applications. However, for small- to
medium-sized biotech firms and bioinformatics departments within pharmaceutical companies,
implementing in-house databases presents a formidable challenge. Part of this challenge is that the
traditional information services department is ill-equipped to deal with the throughput issues that
typically must be addressed by a bioinformatics-compatible network. The typical corporate CIO needs
background education on how to implement gigabit fiber networks dedicated to data storage as well
as high-speed routers and associated network electronics.
Despite the differences between bioinformatics computing and traditional institutional computing, the
process for implementing a high-speed bioinformatics network is identical to that of implementing
any other major network. The major steps in the implementation process are the same, regardless of
whether they are performed by staff in the bioinformatics laboratory or corporate information
services staff. These steps include:
1.
Create a Requirements Specification.
This document includes a high-level description of
the tasks to be supported by the network, such as routing sequencing data from sequencing
machines to analysis workstations and data warehouses, as well as the desired response
times and storage capacities. For example, the requirements specification document may
stipulate the need to support 35 workstations, provide access to storage in excess of 1
terabyte with an access time of less than 50 milliseconds, with tiered password protection,
and secure, high-speed access to the Internet.
2.
Create a Functional Specifications Document.
The functional specifications document
defines, in detail, how the high-level needs outlined in the requirements specification will be
met. This document quantifies many of the qualitative terms in the requirements specification
to the degree that anyone competent in information sciences can determine exactly what
equipment, personnel, and costs will be associated with the project. Once the functional
specifications document has been finalized, the remaining steps are largely straightforward.
3.
Select Hardware.
Assuming the functional specifications document is complete, the next
step is selecting network and workstation electronics and media. Often the functional
specifications document is authored with particular hardware and software in mind, which
further simplifies the selection process.
4.
Select Software.
Again, following the functional specifications document, this step of the
implementation process involves selecting the network operating system, as well as database
publishing software and tools such as PHP, XML, CGI, Java, or JavaScript editors and runtime
systems.
5.
Select Utility.
Software and hardware utilities, such as network monitors and antiviral
utilities, should be defined during the design process, not as an afterthought.
6.
Select Internet Access Service.
Most larger institutions have high-speed Internet access
available throughout their offices. However, bandwidth requirements may necessitate
alternate Internet services, such as supplementing a corporate-wide cable modem service
with a high-speed dedicated line, satellite link, or high-speed microwave link.
Each of the steps in the implementation process requires different levels of expertise with the
bioinformatics requirements, the information technology capabilities, and the likely return on
investment of each approach. As a result, network implementation is necessarily a collaborative
process involving programmers, hardware technicians, vendors, management, and perhaps the
