Stakeholder analysis

From The Business of Energy in U.S. Academic Research Computing
Jump to: navigation, search

A stakeholder is anyone with a interest, or stake, in an activity, enterprise or system. For research computing, the Key Resources page listed a number of resources, of various types, that a research computing organization might have access to. These include equipment, finances, and outcomes. Stakeholders generally refers to people, or groups of people. These people, like other key resources, can have numerous relationships and roles in the research computing organization. Specific individuals can be part of more than one stakeholder group (for example, someone might use a supercomputer, and might also be a collaborator in seeking external funding).

Organizational Stakeholders

Research computing is typically carried out by an organizational unit, or within the auspices of a unit with additional responsibilities for computing and/or research. The former might have labels such as "supercomputing centers," "research computing support services" or "center for innovative technologies." The latter might be part of the university central computing organization, or a service under research administration by a vice chancellor or vice president for research. Stakeholders for research computing units are likely to include:

  • Users: The people who utilize the resources provided by the research computing organization. The extent and type of use can differ substantially, such that a research computing organization might do well to identify the needs and characteristics of each one separately:
  • Light and occasional users: People who make occasional use of research computing technologies, but have only occasional need. This might include persons who use software that is not generally available, have occasional parallel computing jobs, or need larger systems for post-processing or visualization of their research data.
  • Heavy users: It is typical to have an 80/20 distribution in service organizations, where 80% of resources are consumed by 20% of users. For large-scale computation, consumption of CPU hours might be in the millions or tens of millions per year for the larger users, versus hundreds or thousands for occasional users. While heavy users might need less hand-holding and training, they might also be more likely to stress the technologies. A common scenario in supercomputing is for the largest users to dominate resources, such that lighter users need to wait a long time before their jobs can start.
  • Students: People utilizing technologies for class goals. Graduate students might be performing independent research.
  • New users: New users are likely to have additional needs for training or support. They might not be familiar with the nuances of a particular technology or its configuration.
  • Remote users: Some users might be on the same campus as the research computing organization, while others may be remote, or perhaps not even part of the same university or university system. These are likely to have some differing needs for training and support, and might have a different utilization pattern. For example, a remote user who has access to multiple supercomputing centers at different locations might urge your organization to have a similar configuration or policy as others.
  • Personnel: As described in the section on Key Resources. All but the smallest research computing organizations have several people who help to operate and support the technologies. Larger units have their own organizational structures, perhaps with their own managers, budgets, missions, scope of responsibility, and so forth.
  • Collaborators: Generally, collaborators are likely to also be users. Examples include faculty who work with the research computing organization to pursue external funding. Another example is a manager or department head who seeks a strategic alliance, in order to acquire new resources. Collaborators might be internal to the university, or external.
  • Administrators: Research computing organizations exist within an organizational hierarchy, with administrative oversight, reporting, etc. The actual power structure within an organization might not match the organizational chart, so it is incumbent on effective managers to tend to both extant and latent administrative roles.
  • Upstream administrators provide organizational oversight. This is often the Chief Information Officer, Chancellor, Vice Chancellor for Research, or a similar position. These upstream administrators will likely have high-level oversight for the operation and budget for research computing. They might also have day-to-day responsibilities for some aspects of the operation.
  • Lateral administrators: We can loosely define this as any administrator not in the formal hierarchy for the research computing unit. These might have specific roles, which should be understood thoroughly. For example, an administrator in charge of buildings and electricity might play a key role for acquisition of a new supercomputer. A group of administrators at a parallel level to a research computing unit might be important allies.

State & Regional Stakeholders

Higher education institutions are often part of a state or regional group. Public university systems often serve an entire state, where particular campuses might serve a particular geographic area or a set of disciplines (such as a college of engineering, or a medical school). Private universities might have multiple campuses, occasionally in multiple states. For both, there might be regional affiliations and partnerships, for everything from regional Internet (such as SURAnet), to sports (Big 10), to public/private partnerships for commercialization. Some regional stakeholder groups include:

  • State government: Includes oversight and/or control for substantial proportions of state university budgets; engagement in permitting, tax credits, and other activities that could impact research computing data centers, outreach, electrical costs, or commercialization. There are many incidents where state legislative engagement with research computing had a positive impact on its competiveness.
  • Citizenry: Tax payers, voters, residents, at a local, state and regional levels. The largest research computing centers have numerous outreach activities to insure these stakeholders are aware of the positive impact the center has. For any research computing organization, the local citizenry can be instrumental in building a case for mutual success: more and better jobs; bringing high tech industry to the region; competitiveness of students; and scientific discovery with beneficial outcomes.
  • Industry in the region: Most higher education institutions, and essentially all of those with research computing organizations, have active relationships with industry in the area. These relationships help to provide a rich technical job pool, opportunities for public/private partnerships for research & development, and, in some cases, opportunities for selling research computing services to industry.
  • Agencies: State agencies, local branches of federal agencies, non-profit organizations, and city/county agencies may all have similar relationships to those mentioned for state government and industry. Some agencies might benefit directly from the research computing organization's activities. This could be via outreach projects, a high-tech jobs or training program, or mission-focused research (for example, a supercomputer might be used to model the local aquifer).

National and International Stakeholders

  • Industry/academic/professional associations: These can be advocates for shared interests. They can also be sources of examples for policies or best practices within research computing. In some examples, including CASC, these assist in information and advocacy to federal lawmakers.
  • Federal government: Especially through influence over federal agencies, the federal government sets policy and funding levels for investment in research. This includes research computing. Major strides in funding have often accompanied federal government initiatives, such as occurred in 2012 with the federal Big Data launch.
  • Industry in high technology: High technology industry is in a constant state of growth, and some of that is due to productive relationships with research computing centers. Leadership-class supercomputers leverage investments by federal agencies and others to develop, implement and evaluate new technologies that might eventually become mainstream. Good relationships with leading vendors are helpful for bringing new technologies into research computing organizations, and large vendors are proactive about state and federal funding opportunities.
  • International organizations: These are often supersets or conglomerations of national organizations, such as professional associations (IEEE is one of the largest), standards organizations such as IETF and OGF, and multinational government partnerships (such as the University of the Arctic). In some cases, they have a specific mission to foster international developments in research computing. The National Science Foundation is typically the US lead agency for academic ventures of this sort.