As for this assignment, we were tasked by our professor to interview our university network specialist and ask how the various parts of the system communicate with each other throughout the university. We were also asked by our professor that, given the chance to redesign the existing setup, we enumerate and discuss our key points for an effective and efficient network ideal for the university.
I think that almost everyone (or everyone here) has the same set of answers from our Network Administrator, and that’s because we also have the same set of questions for him. Not all questions have been answered though (as of this time).
Here are our questions and his answers to those questions. (“US” here represents the students)
US: In system development, how various parts of the system communicate with each other throughout the university? In what way?
SIR REYES: “Regarding System Development, the best person to ask is our University Programmers, Mr. Fortich and Dr. Mercado.”
US: What are the components involved in the system(s) in the university? (hardware, software, technology, etc.)
SIR REYES:” I am not in the right position to discuss the details of the software components used as there are other assigned personnel for such job. However, talking about hardware component and technology used, basically I, assigned as the network administrator, is entrusted to maintain our different servers to run 24/7. Currently, we have our Web Server hosted here in our University in our HP ProLiant ML350 Server. Its an old but stable server set-up here in our Networks Office and has been active since Engr. Val A. Quimno , not yet a dean, was appointed as the Network Administrator. The said server has the following specification:
· Intel Xeon 3.0 GHz, 3.2 GHz, or 3.4 GHz processors (dual processor capability) with 1MB level 2 cache standard. Processors include support for Hyper-Threading and Extended Memory 64 Technology (EM64T)
· Intel® E7520 chipset
· 800-MHz Front Side Bus
· Integrated Dual Channel Ultra320 SCSI Adapter
· Smart Array 641 Controller (standard in Array Models only)
· NC7761 PCI Gigabit NIC (embedded)
· Up to 1 GB of PC2700 DDR SDRAM with Advanced ECC capabilities (Expandable to 8 GB)
· Six expansion slots: one 64-bit/133-MHz PCI-X, two 64-bit/100-MHz PCI-X, one 64-bit/66-MHz PCI-X, one x4 PCI-Express, and one x8 PCI-Express
· New HP Power Regulator for ProLiant delivering server level, policy based power management with industry leading energy efficiency and savings on system power and cooling costs
· Three USB ports: 1 front, 1 internal, 1 rear
· Support for Ultra320 SCSI hard drives (six hot plug or four non-hot plug drives supported standard, model dependent)
· Internalstorage capacity of up to 1.8TB; 2.4TB with optional 2-bay hot plug SCSI drive
· 725W Hot-Plug Power Supply (standard, most models); optional 725W Hot-Pluggable Redundant Power Supply (1+1) available. Non hot plug SCSI models include a 460W non-hot plug power supply.
· Tool-free chassis entry and component access
· Support for ROM based setup utility (RBSU) and redundant ROM
· Systems Insight Manager, SmartStart, and Automatic Server Recovery 2 (ASR-2) included
· Protected by HP Services and a worldwide network of resellers and service providers. Three-year Next Business Day, on-site limited global warranty. Certain restrictions and exclusions apply. Pre-Failure Notification on processors, memory, and SCSI hard drives.
Aside from it, our mail server running under Compaq Proliant ML330 Server, our oldest server, is also hosted here in our Networks Office. Together with other Servers, such as Proxy and Enrollment Servers, both proxy and our enrollment servers are running in a microcomputer/personal computers but with higher specifications to act as servers.”
SIR REYES: “All Servers are connected in a shared medium grouped as one subnetwork. In general, our network follows the extended star topology which is connected to a DUAL WAN Router that serves as the load balancer between our two Internet Service Providers. All other workstations are grouped into different subnetworks as in star topology branching out from our servers subnetwork as in extended star topology. At present, we are making use of class C IP Address for private IP address assignments. Other workstations IP assignments are configured statically (example: laboratories) while others are Dynamic (example: offices). All workstations are connected via our proxy servers that do some basic filtering/firewall to control users’ access to the internet aside from router filtering/firewall management. So, whenever any workstation has to connect to the internet, it has to pass through software and hardware based firewall.”
US: What are the processes involved in the communication (each system to other systems)?
SIR REYES: “As mentioned above, in item 3, all workstations are connected via a proxy server. It means that whenever a workstation is turned on, it requests for an IP address from the proxy server (for dynamically configured IP address) and connect to the network after IP address is acquired. As connection is established, each system can now communicate and share resources within the same subnetwork and to server following the concepts discuss in your Computer Network Class.”
SIR REYES: “Basically, our servers are expected to be in good condition since it is required to be up 24/7. Daily, during my vacant period, monitoring on the servers are observed that includes checking logs, checking hardware performance such as CPU health, etc. If problems are observed, remedies are then and then applied. Once in a week, regular overall checkup is observed as preventive maintenance to ensure not to experience longer downtime if possible.”
US: Does the system follow a specific standard? Explain Please.
SIR REYES: “As I was appointed as the Network Administrator, everything was already in place except for some minor changes. Basically, different networking standards was already observed such as cabling standards, TIA/EIA 568A-B, different IEEE standards as discussed in your Computer Networks Subject, etc.”
US: How is the security of the system? Are there any vulnerabilities? Risks? Corresponding mitigation techniques? Access control?
SIR REYES: “As I have mentioned, we have implemented both software and hardware based filtering/firewall. Basically, Risks or vulnerabilities and different mitigation techniques were considered to increase security in our network. Aside from filtering/firewall, constant monitoring on networks activity also increases the security of the system.”
US: Are there any interference? During what (most) times do these occur? Explain their effects especially with regards to the business of the university?
SIR REYES: “Major Interferences are normally encountered as an effect of unforeseen and beyond our control events such as black outs, and the like. The said interference would of course affect University’s day-to-day businesses for obviously this will paralyze all our activities that rely on electricity and further this might cause damage on our network devices, etc. that may later be the reason for longer downtime. Problems encountered by our providers such as connection to the National/International Gateway also affect University’s business such as correlating to University’s Business Partners outside and within the country.”
And that’s about it. There were other questions thrown for our network specialist, but sadly these were not answered.
As for my answers if we were given the chance to redesign the existing setup of the university’s network environment…
First of all, we have to plan for the re-engineering of the current setup of the network environment here in the university. I read some stuff from the net about planning and here is what I got.
Planning is a process, not a product
Some colleges use formal methodologies for planning, while others do planning "on the fly." Whatever the procedure used on your campus, it is important to stress that planning is a process, not a product. This is one of those "trite but true" statements about planning. It may be especially true (or especially trite) when applied to network planning. But a network is an organic entity that will continually evolve, grow, and mutate. Thus colleges need appropriate feedback mechanisms, organizational structures, and planning processes for measuring growth and handling change in the network.
Network planning must be mission driven
The network planning activity must be connected to the strategic goals and directions of the institution. What is the primary mission of the institution, and how does this affect networking? For example, if residentiality is a key part of the college’s mission, how does networking relate to that? As colleges develop strategic plans for the institution, or master plans for facilities, it is very important to include the campus network at appropriate places in the overall planning.
Network planning is planning for the new ICE Age, in which Information, Communication, and Entertainment converge
It is important to understand that network planning really means planning a whole new environment that will encompass information, communication, and entertainment on the campus. The network will provide a powerful, dynamic, new channel of access to the world of information. It will also enhance and fundamentally alter the communication mechanisms on campus. And whether we want it or not, the network will provide a wide range of new entertainment opportunities. This new environment will encompass or touch all the information technology areas (e.g., computing, library automation, audio, video and multimedia technologies), which will challenge, compete with, and sometimes replace existing modes of campus communications, information access, administrative work, power hierarchies, and budgeting systems.
While change has always been a characteristic of technology, the rate of change is increasing. New technologies and combinations of technologies crash into our lives almost constantly. It is difficult to even know about all the new technologies, much less understand the implications of them for our campuses. But campus network builders have to be both scouts and wagon masters of the information age, so it is important to find ways to step back, reflect, and observe the changes taking place. We need this to get a clear perspective, perceive subtle directions and trends, and respond to new imperatives before they become crises.
Planning the ideal network
Physical dimension. The ideal campus network is an information channel that reaches every place on campus where "knowledge workers" live and work, including offices, classrooms, laboratories, studios, student residences, student activity areas, and so forth. It includes a physical infrastructure that consists of high-grade copper and fiber cable; junction and termination boxes; communications devices such as fiber hubs, bridges, routers, terminal servers; and wiring closets to house equipment and termination panels. The ideal campus network provides a seamless interface to on-campus sub-nets and to off-campus locations and resources, such as faculty homes, metropolitan and regional networks, and the Internet. The ideal campus network has physical components that meet defined institutional standards, provide for modularity and expandability, and are well documented and mapped.
Protocol dimension. The ideal campus network handles multiple network protocols, such as TCP/IP, AppleTalk, Netware, etc. Therefore it should not have highly proprietary characteristics that preclude use of other protocols. The ideal network provides a seamless interface between protocols used on sub-nets and meets well-defined institutional standards for network connections and protocols.
Management dimension. On the ideal campus network, management activities are invisible to users. Network growth, while constant, is managed without disruption to users. The ideal network management structure includes appropriate staffing, as well as budgeting, control, and security systems. The institution with an ideal campus network has a funding program that covers continual growth of the network and replacement of functionally obsolete equipment.
Application dimension. The ideal campus network provides easy access from any connection point to all information pools, including the global Internet, library materials, specialized departmental resources, non-print media collections, and institutional databases. The network provides a variety of integrated information resources via a campuswide information system (CWIS). The ideal campus network incorporates a seamless electronic mail system with a common user interface to all members of the institutional family, which may include off-campus constituencies, and provides easy sharing of electronic resources (data, text, images, sound, video) across the network. All members of the campus community find it easy to use shareable computing hardware and software resources such printers, scanners, statistical packages, programming languages, and databases; everyone has full access to Internet applications and information resources; and all applications are well documented and publicized.
Cultural dimension. Ideally, all faculty, staff, and students use the network fluently as a natural and integral part of their communications and information exchange activities. On a campus with an ideal network, students use the network as an intrinsic part of their campus life, faculty actively seek to use the network in new and creative ways to enhance teaching and learning, and administrators and staff routinely use the network to improve operations and reengineer archaic administrative systems. The network provides a unifying concept for campuswide integration of information technologies, resources, and services. The institution considers the network a strategic asset, and is committed to supporting the network as a vital strategic resource.
Planning and building campus networks often proceed together in an iterative fashion. Even if we share a common vision of the "ideal" network environment, each campus network will have to be tailored to that campus. The sidebar opposite offers some general guidelines for building an ideal network.
Funding issues
Funding is a central issue in planning and building networks. The financial part of network planning should include a strategy based on capital funds, support funds, and maintenance funds.
Capital funds are generally used for building the basic network infrastructure. Every campus network is different, so it is very difficult to provide a general formula or model to accurately predict the design and development costs. For example, a campus with a network of steam tunnels will find it much cheaper to install conduit and fiber than one that has to dig up or punch under city, county, or state roads.
Support funds, often included in the institution’s operating budget, are required for the ongoing support of network components and services. Examples of such expenses include personnel costs, license fees, Internet line charges, and fees for regional providers of Internet connections.
Maintenance funds will be required to replace damaged, worn out, or functionally obsolete networking equipment. Administrators must understand that most information technology items are just like other essential utilities and "consumables" (for example, water, electricity, and toilet paper), and as such, must be funded as part of the institution’s ongoing operating budget. Keeping an equipment inventory, with the expected replacement cost and the expected lifetime for each item, is necessary so yearly costs can be predicted. For simplicity, if a component is expected to last N years, then each year (100/N) percent of the replacement cost should be deposited in a depreciation reserve fund. And the network management staff should have access to this reserve account on an "as needed" basis.
Staffing and human resources issues
What are the staffing implications of networking? What are the new responsibilities related to the development and management of a full-campus network? Some areas where staff time will be required include:
User support. Faculty, staff, and students will need all the types of support that computing requires (consulting, training, documentation).
Software support. Network operating systems will need regular management, maintenance, and upgrading. Applications are increasing dramatically in both quantity and variety.
Technical support. Someone will have to run cable, make and maintain network connections, trouble-shoot problems, install routers, and so forth. This kind of service could be outsourced or could be done by college employees.
External relations. Someone will have to work with vendors and other external partners, represent the institution in maintaining an Internet connection, talk to parents and alumni who want access to the campus network, and perhaps help with external fund-raising.
Policy issues. Someone will have to coordinate the development of campus policies and procedures to address the many questions that will arise when access is provided to electronic information about its proper use, ownership, authorization, control, accuracy, security, privacy, and so forth.
Building the Ideal Campus Network
Aim for the ideal
Design the network to reach everywhere, including all offices, classrooms, labs, studios, student residence areas, and auxiliary enterprises, anticipating new buildings, campus reconstruction, and building renovations. Don’t believe it if someone tells you that a particular building will never need to be connected to the network. Include network support facilities such as wiring closets, security systems, power conditioning, and lightning protection in the design. Wire the student residences for data, voice, and video.
Get the power and build alliances
Some individual needs to be designated as the final authority on all campus network development to coordinate planning and construction, ensure necessary access to all campus buildings and spaces, push the creation of standards, and build campus alliances with important stakeholders. Such authority should reside in a position that reports at a strategic level in the institution, which usually means to either the chief executive officer or someone who reports to the CEO. The responsible person needs to have a broad institutional perspective and be a good communicator, consensus builder, and pragmatic visionary.
Develop and enforce standards
Clear institutional standards for all components of the physical network, as well as the software and hardware that will provide network services, are critical. For example, at the construction stage, standards must be defined to proceed with the basic steps of writing RFPs, comparing bids, making purchases, and testing wiring. While diversity is valued in many aspects of collegiate life, technological diversity is expensive in terms of time, money, and effort. Standards make it possible to maximize the use of human, financial, and physical resources, to train and deploy technical staff, and to support users. These support issues have a direct impact on the quality of the technology environment as well as the basic teaching mission.
Accept that too much is never enough
Make the network design as flexible as possible. The "giga-world" has arrived, and bandwidth, speed, and quality-of-service demands will continue to grow. During construction, put in extra conduit wherever and whenever possible, allowing for flexible cabling strategies such as "blown fiber." Ample fiber should be used between buildings, with either direct fiber runs or Category Five unshielded twisted pair copper cable to the desktop within buildings. Assume that all network electronics (e.g., routers, hubs, bridges, and switches) will have to be replaced regularly, generally about every two to three years. Using adequate fiber and high-quality UTP wire should make it possible to move up to any speeds likely to be necessary in the foreseeable future.
Take the long view
Technologies such as the World Wide Web, wireless communications, networked multimedia systems, and so forth are having a dramatic impact on all educational institutions. The next iteration of the global Internet, while not yet a commodity system, will bring affordable high-performance networking to all colleges within a few years. The best thing that small colleges can do to get ready for the next generation of networking is to build a strong, supportable, well-managed campus network today.
Don’t forget to remember history
Record the details. Document and map everything. Without detailed maps, it is amazingly hard to remember where underground conduit runs a year or two after it has been buried. It is difficult but worth trying to get the contractors to provide "as built" diagrams after a construction job is finished. Keep copies of all RFPs, proposals, plans, diagrams, and blueprints. And keep them in a logical, central place, not as head files. It may be worthwhile to invest in a CAD package and scanner to help record the construction details.
Help it grow
Campus network use never declines. Even as some types of usage mature and level out, other network-based applications arise and consume new chunks of the network resources. Growing the network to keep up with the demands for more connections, more speed, and more bandwidth requires continual funding. Justifying regular funding may require cost/benefit arguments based on network use, so it is useful to develop a set of metrics for network-related activity to help administrators understand the role and impact of the network on campus life. Measure such things as live network connections and network load/traffic, number of people using the network, number of network connects to the online catalog, number of hits on the Web site, and so forth.
MY VIEWS:
I, myself agree on what the article had said about campus networking. At first, we have to plan on what to do. As my specific knowledge on networking is not that sufficient enough to answer some questions regarding technical issues, I could not specify my answers on that. But for me, I would really choose the best topology, the most suitable equipment, and the most suitable people for the job so that there would be less downtime and surfing the net in school would be faster.
Reference: http://net.educause.edu/ir/library/html/cem/cem99/cem9916.html
No comments:
Post a Comment