Question

A local hospital asks you to help improve its networks fault tolerance. The hospitals network carries critical patient care data in real time from both a mainframe host and several servers to workstations in operating rooms, doctors offices, the billing office, teaching labs, and remote clinics across the region. Of course, all of the data transferred is highly confidential and must not be lost or accessed by unauthorized personnel. Specifically, the network is configured as follows:

Six hundred workstations are connected to five shared servers that run Solaris. Fifty of these workstations serve as training computers in medical school classrooms. Two hundred workstations sit in doctors offices and are used to view and update patient records, submit accounting information, and so on. Twenty workstations are used in operating rooms to perform imaging and for accessing data in real time. The remaining workstations are used by administrative staff.

The clients are connected in a mostly switched, star-wired bus network using Ethernet 100Base-T technology.

An Internet gateway supports e-mail, online medical searches, and VPN communications with four remote clinics. The Internet connection is a T3link to a local ISP.

A firewall prevents unauthorized access from the T3 connection into the hospitals network.

The hospitals IT director asked you to identify the critical points of failure in her network and to suggest how she might eliminate them. On a sheet of paper, draw a logical diagram of the network and identify the single points of failure, then recommend which points of failure should be addressed to increase availability and how to achieve this goal.

Case Project 2

Unfortunately, the solution you provided for the hospital was rejected by the board of directors because it was too expensive. How would you determine where to cut costs in the proposal? Consider which fault-tolerant options are most critical and also which are expensive but less critical. What questions should you ask the IT director that will better enable you to prioritize the fault-tolerance features youve recommended? What points of failure do you suggest absolutely must be addressed with redundancy?

Case Project 3

Your second proposal, with its reduced cost, was accepted by the board of directors. Now, the hospitals IT director has asked you to help develop a disaster recovery plan. Based on what you have learned about the hospitals topology, usage patterns, and current fault-tolerance measures, outline a disaster recovery plan for the hospital. Your plan should specifically address provisions for data, equipment, and connectivity related to the network. Explain how functionality and data will be restored and what staff should be involved in the post disaster recovery.

Answer 1

With the information given there are a number of single points of failure. The Star topology in my opinion would be the largest, followed by the single T3, and an Internet Gateway that seems to support three different business critical items (email, VPN connections, and online medical searches).Although a single firewall is not usually recommended you can make sure that if the firewall dies on you that no traffic comes into the business. On the other hand, no traffic will be able to leave, either.

Can you elaborate more on why the star topology would be the most critical and the T3 link as well?

The star topology means that if one link in the star goes down near the center of the topology that all the links downstream would be dead.

A mesh topology would be the best, can quickly become expensive if every device has two connections.I'm sorry, I'm thinking hub and spoke.A star topology would work.

Ideally you would want dual T3's, with access to the internet from different providers.If one goes down, or if one provider has large problems the other link will be available.I will be able to answer any further questions in about half an hour.

so the star topology would work.. but it is still a point of failure?what are the recommendations for those points of failure? how should those be addressed to increase availability?
I'm sorry, I was thinking of a different topology when I listed it as a single point of failure.That would not be a single point of failure.There would ideally be two T3's from different providers bundled together for ease of use by a router using EtherChannel or a similar protocol. Should one of them experience trouble there would be no human intervention to keep the connection up.

so the points of failure are the T3 connection and you are recommending two T3's so one can serve as a backup what about the firewall?

The internet gateway would need to be duplicated. If all three items (email, VPN, and online medical searches) are housed on the gateway it would be best to duplicate these on the second gateway.The firewall would be best to duplicate also.You could use a DNS round robin to share the load between the internet gateways and firewalls, or many other High Availability protocols.