TECHNICAL ENGINEER EXAMINATION (NETWORK)(AFTERNOON, PART 2)

Chia sẻ: Tran Le Kim Yen Tran Le Kim Yen | Ngày: | Loại File: PDF | Số trang:0

Thêm vào BST

Báo xấu

42
lượt xem 4
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

Which of the following is a correct statement in regard to a preemptive system in OS process control? a) Each process can manage system resources autonomously, so the preemptive system is well-suited to multi-programming.

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: TECHNICAL ENGINEER EXAMINATION (NETWORK)(AFTERNOON, PART 2)

2001 Autumn Technical Engineer Examination (Network)(Afternoon Part 2) Questions must be answered in accordance with the following: Q1 to Q2 Question Nos. Question Selection Select one of the above two Examination Time 14:30-16:30 (120 minutes) Instructions: 1. Use an HB pencil. If you need to change an answer, erase your previous answer completely and neatly. Wipe away any eraser debris. 2. Mark your answers in accordance with the instructions below. (1)Examinee Number Write your examinee number in the space provided, and mark the appropriate space below each digit. (2)Date of Birth Write your date of birth (in numbers) exactly as it is printed on your examination admission card, and mark the appropriate space below each digit. (3)Answers In the Selection column, circle the numbers of the questions that you are choosing to do. Any question that is not circled will not be scored. If you circle all four questions, only the first three will be scored. (4) Write each answer in the space specified for that question. (5) Write your answers clearly and neatly. Answer that is difficult to read will receive a lower score. Do not open the exam booklet until instructed to do so. Inquiries about the exam questions will not be answered. 1
2
[List of Question Contents] Question Theme Questions No. [Description] Sub-Question 1 (1) Basic knowledge regarding IP-VPN service (2) Reason why leased lines are used for access lines Sub-Question 2 (1) Names of sites which require the study of the band width of the leased lines (2) Effect of adding sites on the existing network Sub-Question 3 (1) Problems resolved by installing a server for the integrated application program in the main office (2) Reason why increased traffic between the plant and main Re-construction of a office affects other sites Q1 network using IP- Sub-Question 4 VPN service (1) Basic knowledge regarding connections for data transfer (2) Port numbers sent by the PORT command (3) Details of changes made to the IP address of an ISDN router (4) Reason why communication was disconnected by an ISDN router (5) Information useful in troubleshooting when IPsec is not used Sub-Question 5 (1) Reliability problems solved by the new network (2) Reason why communication with the main office is not possible if the IPsec function is implemented using PCs in regional sales offices (3) Features and associated reasons behind IP address planning when using IP-VPN service [Description] Sub-Question 1 Basic knowledge regarding duplexing ISPs Sub-Question 2 (1) Reason why the reply packet is returned to the same firewall (2) Hardware that should be checked using ping in order to detect routing failures Sub-Question 3 (1) Basic knowledge regarding duplexing mail systems (2) Reason why implementing processing to prevent illegal relays is simple Sub-Question 4 (1) Basic knowledge regarding RAID (2) Reason why a load balancing device is not used for distribution to an application server Safety measures for (3) Zone information managed by DNS server 2 Q2 a system connected Sub-Question 5 to the Internet (1) Basic knowledge regarding use of housing services offered by an Internet data center (2) Role of a UPS when switching to an in-house power generator (3) Tasks that should be indicated in an operation management manual in order to operate an electronic commerce system (4) New issues that Company Y should consider to prevent system failure when using the housing service of an Internet data center [Illustration] Sub-Question 5 (1) Completing the configuration of an electronic commerce system when using the housing service of an Internet data center 3
Q1. Read the following description of re-constructing a network using IP-VPN service and answer Sub-Questions 1 through 5. Company A mainly sells customized PCs to corporations. Since the scope of the business is rather large, Company A has a regional sales offices in charge of each region and branch offices that oversee them. For their salesmen and designers, corporate customers order PCs with optimum specifications to do their jobs. When a salesman receives an order, he sends to the plant the customization specifications as requested by the customer. A PC is assembled according to the customization specifications. The assembled PC is sent to a distribution center and shipped on the specified day of delivery to the customer who placed the order. [Overview of the Previous System] Company A used to run separate business application programs (hereafter referred to as “distributed business APs”) on servers located at its branch offices and plants. The distributed business AP on each server was used with a corresponding client application program (hereafter referred to as a “terminal AP”) which ran on PCs. They were connected via TCP/IP communication lines. The distributed business APs at branch offices were used for order entry and business activity reports, while the distributed business AP at plants was used for production control. In addition, the terminal AP used by salesmen to make business activity reports was accessing through FTP to the distributed business AP used to make business activity reports which ran on servers at branch offices. [Background Behind New System Development] A year ago, the Planning Department of Company A began sales of PCs to individuals using a Web page-based online sales system installed in the main office. Even in the case of PCs sold to individuals, there were many cases where PCs needed to be customized to meet the requirements of individual customers. In the case of sales to individuals, customers wanted to be able to access the status of their order at any time from the PC order placement to its delivery. This service is called a “tracking service”. Salesmen also wanted this service in order to quickly respond to inquiries about delivery and so forth from corporate customers. 4
Although the distributed business AP was developed by the Information System Department at the main office, there were many problems because many additional functions had been repeatedly added. Maintenance personnel were used to take care of problems at branch offices and plants, but many times they could not solve problems and the Information System Department had to take care of it. This was a hindrance to the development work being performed by Information System Department. [Overview of the Current System] Six months ago, Company A got rid of the distributed business APs and distributed servers that it had been using, and began running a newly developed business application program (hereafter referred to as an “integrated business AP”) on a new server located in the main office. The integrated business AP included a function for linking with the nonstore sales system and a function for tracking service, while also implementing all the business functions of the old distributed business APs. The new server could be used from a PC using TCP/IP. It was decided to aim at quick development of the current system and to continue the use of the previous system. Fig. 1 shows the configuration of the current system network. Regional Sales Office 1 Internet PC ISDN … Router Leased PC Main Line Router RAS Router W FW Office … ISDN Leased PC … Branch … PC PC S PC Line Office PC ISDN … Plant Research Leased Line Router Router Router Lab PC Leased PC … PC … Line PC PC Regional Sales Office 5 FW: Firewall (details of configuration of Distribu- firewall omitted) Router tion Center RAS: Remote Access Server W: Web server for nonstore sales system PC … PC S: Server for the integrated business AP Note: Dotted lines indicate separate sites. Fig. 1 Configuration of Current Network 5
Salesmen in regional sales offices also have demands. Business activity reports are made by accessing the integrated business AP using a terminal AP for making such reports as in the past. Since terminal APs other than this have been abolished with the operation of the integrated business AP, it was decided to use the integrated business AP using PC browsers. A salesman can therefore use a browser to find out the status of a PC order at any time up to its delivery. [Reconstruction of the Current Network] Business at Company A has increased steadily using the integrated business AP. With this increased business, more working hours are spent referencing the specifications and design documents. The traffic between the plant and the main office has seen particularly dramatic growth. This has caused longer response times at multiple sites on the network and is hindering business. However, the communication bandwidth between the main office and branch offices is sufficient and there are no problems here. The Information Systems Department was assigned to study the reconstruction of the current network for better network reliability and expandability to handle increased traffic and the addition of new sites in the future. The Information Systems Department, with Mr. T as a leader, has collected the requirements of the new network and presented its findings to a communication service provider. As a result, Mr. K, an engineer working for the communication service provider, has proposed IP-VPN service using MPLS (Multi-Protocol Label Switching). It was decided to study with him the suitability of IP-VPN service for the new network. Fig. 2 shows the configuration of the new network using IP-VPN service as proposed by Mr. K. Regional Sales Office 1 Internet Branch Office Leased Leased PC Main Line Line ISDN Router Router W FW Office Router PC PC PC S PC PC ISDN IP- Leased Line Router Research Plant Leased Line Router Lab VPN PC ISDN Router PC PC PC PC PC Distribu- Regional Sales Router tion center Office 5 PC PC Fig 2. Configuration of New Network 6
The following is a conversation between Mr. T and Mr. K. Mr. T: First, please tell me about the packet transfer method used with IP-VPN. Mr. K: The router connected to the leased lines in Fig. 2 are called “customer edge routers”. (Hereafter referred to as CERs.) When using IP-VPN service, the communication service provider’s provider edge routers (hereafter referred to as PERs) are connected to customer CERs via leased lines having the required bandwidth. This leased line is called an “access line”. An IP packet arriving at a PER from a CER is given a/an at the PER a based on its destination address. Inside the IP-VPN network, routing between the sending PER and destination PER is performed based on is removed by a a . at the transfer destination, restored to a regular , and transferred to b c . d Mr. T: Inside the IP-VPN network, packets having a different format than IP packets are transferred, right? Mr. K: Yes, that’s right. Mr. T: Can security be achieved when using IP-VPN service? Mr. K: Of course it can. At the sending PER, it is possible to know which IP packets came from which customer. This does not arrive at the CERs of other customers. In other words, the sending PER identifies the sending customer, and determines the destination PER according to the destination IP address in the received IP packet. If the sending customer is different, that IP packet is transferred to CER even if the destination IP address e in the IP packet received from the customer exactly matches that of the sending PER. This allows security equivalent to communications to be achieved using conventional leased lines. Mr. T: Please tell me about the case of future expansion stated in your proposal. Mr. K: For example, imagine that you are going to establish a new distribution center. In your current network configuration, this means connecting the new distribution center to the plant using a leased line. In this case, it is also necessary to study the bandwidth of existing leased lines between and and between and . In f g g h contrast, in the new network configuration being proposed, expansion will be easy because the effect on the existing network of adding sites can be just by studying the i bandwidth of existing access lines between and IP-VPN. f Mr. T further continued his investigation on the assumption that a new network would be configured using IP-VPN service because it allows for communication security and can be done at low cost. 7
[Connecting the Regional Sales Offices to the Main Office] Mr. K’s proposal was that regional sales offices and the main office be connected over the Internet. Mr. T investigated the method of the connection with Mr. K. Mr. K: Connections to regional sales offices shall be made with the main office which has the Information Systems Department. Since the main office and each regional sales office are physically separated, connections which use ISDN in the current network will be switched to connections that use the Internet. Mr. T: Although I think it is appropriate to connect regional sales offices to the main office, the proposals from other communication service providers suggest using IP-VPN service at regional sales offices as well. Why doesn’t the proposal from your company suggest IP- VPN service be used at the regional sales offices? Mr. K: We feel that the frequency of use of the network by regional sales offices is low. We therefore thought that IP-VPN service was inappropriate because it is not very cost effective to use leased lines as the access lines with regional sales offices. Mr. T: Can anything besides leased lines be used as the access lines? Mr. K: Nothing can be used except leased lines. Mr. T: Although security is achieved under IP-VPN, I’m very concerned about security when communications are made over the Internet. Is there anything we can do? Mr. K: Security for communications over the Internet between the regional offices and main office can be achieved at the IP layer by using IPsec to safely transfer IP packets. When using IPsec, the sender encrypts IP packets and the receiver decrypts encrypted IP packets. Figure 3 shows a basic overview of using IPsec packets as currently being proposed. IP packet before encryption ESP Authentica- Supplemental New IP header ESP header IP header TCP header TCP data i ESP data tion data Target of encryption Target of identification ESP: Encapsulating Security Payload Fig. 3 Overview of IPsec Packets Used at Company A 8
Mr. K: The IP packet before encryption and the newly added supplemental ESP data are encrypted. The ESP header, which differs from a TCP header, and encrypted data are the target of falsification detection. Mr. T: Are there any problems with reduced communication throughput due to encryption overhead or Internet congestion? Mr. K: The main business being conducted by regional sales offices is producing business activity reports. The terminal AP used for making business activity reports uses FTP to connect to the integrated business AP and download a report template. Then report data created by a salesman is sent to the integrated AP. Since the amount of data sent with FTP is small, we think that it is not a big problem. Mr. T performed a file transfer test using IPsec and FTP over the Internet. [Connection Test between Regional Sales Offices and the Main Office] First, Mr. T made preparations to perform a connection test using FTP over the Internet between each regional sales office and the main office. Although the FTP server used for the test possessed an IPsec function, in preparation for trouble analysis, the IPsec function was not used. Company A uses a private, in-house IP address. In order to connect to the Internet, Mr. T set the packet filter for ISDN routers used on the current network for use under an Internet connection. The table gives an excerpt of the contents of packet filters for ISDN routers used for the connection test. Table: Contents of Packet Filters of ISDN Routers Used in the Connection Test (Excerpt) Sender’s Destination IP Sender's port Destination Communicati Direction SYN bit ACK bit IP address address no. port no. on operation 1,024 or Out-bound Internal External On Off 25 Connect greater 1,024 or Out-bound Internal External On Off 110 Connect greater In-bound External Internal On Off Any Any Disconnect 1,024 or In-bound External Internal Off On 80 Connect more Out-bound Internal Internal On Off Any Any Disconnect 9
In addition, it was possible to use the Internet from multiple PCs which possessed a private IP address by assigning a single global IP address to each of the ISDN routers used in the test. Mr. T downloaded files from an FTP server located in the DMZ of the firewall from a regional sales office using a browser. Next, he tried downloading the files using the terminal AP for making business activity reports. However, he could not download files from the FTP server using the terminal AP for making business activity reports. Mr. T reported the results of this test to Mr. K and decided to find out the causes of problems and their solutions. Mr. T: I was able to download files when using the browser in a regional sales office. However, download failed when attempting to use the terminal AP for making business activity reports. Here is the packet monitoring data between the FTP server and the FTP client obtained from the DMZ of the firewall. Looking at this data, it appears that the TCP connection from the regional sales office was disconnected, but I don't understand the cause. Mr. K: Judging from the monitoring results, the TCP connection was disconnected by the ISDN router. I will explain using Fig. 4, which shows an overview of FTP active mode. FTP server FTP client P: 21 P: 20 P: 3201 P: 3200 Establish connection for control Transfer PORT command Transfer RETR command Establish connection for data transfer Transfer files Connection for data transfer disconnected Connection for control disconnected Note: “P” indicates the Port No. Fig. 4 Overview of FTP Active Mode 10
Mr. K: First, the FTP client requests that a connection for control be established. The IP address and Port No. of the FTP client are sent to the FTP server by the PORT command using the connection for control. This information is used to establish a connection for data transfer. The request for the establishment of a connection for data transfer is made by . j FTP active mode is used based on packet monitoring data for the terminal AP used for making business activity reports. Mr. T: Why can’t files be transferred when the terminal AP for making business activity reports is using FTP active mode? Mr. K: That’s because the packet filtering setting of the ISDN router does not correspond to FTP active mode. However, if router settings are made to correspond with FTP active mode, there is a possibility that the security at regional sales offices will decrease. Mr. T: Since the FTP client is separated from the current network under the current environment, I don’t think that in-house security is a problem. I'm going to perform a connection test using different ISDN router settings. Mr. T confirmed that it was possible to download files using the terminal AP for business activity reports in an environment which does not use the IPsec function. According to the proposal made by Mr. K, existing ISDN routers were to be changed to models with an IPsec function. Mr. T confirmed that it was possible to download files using the terminal AP for business activity reports in an environment that uses the IPsec function when using ISDN routers that include the IPsec function. In addition, he also confirmed that communication throughput did not decrease even when the IPsec function was added. [Investigation of the IPsec Function] In order to reduce the costs, Mr. T proposed using the IPsec function by software processing on a PC installed in the regional sales offices, rather than switching existing ISDN routers to models with the IPsec function and discussed this possibility with Mr. K. 11
Mr. T: Although plans call for switching to ISDN routers with an IPsec function, recent PCs possess an IPsec function that is implemented in software. I would like to consider the possibility of implementing IPsec also. Mr. K: The main difference is in the way in which private IP addresses used inside the company are converted to global IP addresses used for regional sales offices. There is only one global IP address assigned when using a dialup connection to the Internet. Mr. T: ISDN routers have an IP address conversion function, don’t they? Mr. K: Yes, they do. The IP address conversion system used is called “IP masquerade”. Fig. 5 shows a summary of this function. Client Server Ad: Global IP address Ad: Global IP address As: Global IP address As: Private IP address Pd : 80, Ps : 1050 Pd : 80, Ps : 3050 ISDN router Ad: Private IP address Ad: Global IP address As: Global IP address As: Global IP address Pd : 1050, Ps : 80 Pd : 3050, Ps : 80 Note: “Ad” indicates the destination IP address, while “As” indicates the sender’s IP address. “Pd” indicates the destination port no., while “Ps” indicates the sender’s port no. Fig. 5 Overview of the IP Masquerade Function Mr. K: When using IP masquerade, the private IP address of the IP packet received from the client is converted into the global IP address of the ISDN router. At this time, the sender’s port no. is also converted to a number selected by the ISDN router. An IP packet is then sent from the server to the converted global IP address and port no. The ISDN router converts the destination IP address in the IP packet received from the server into the original private IP address, and also restores the port no. For this reason, there is a problem with the order in which the IP masquerade function and IPsec function are applied and it is therefore difficult to use a PC’s IPsec function. Mr. T: I understand. We will use ISDN routers with an IPsec function. Company A began transition to its new network based on the results of investigation which had been made so far, and proceeded to create a concrete transition plan. 12
Sub-Question 1 Answer the following questions regarding the IP-VPN service to be used with the new network. (1) Fill in the appropriate words/phrases for boxes through in the a e text. (2) In 80 words or less, explain the reason why leased lines are used for access lines from a security perspective when PERs are used. Sub-Question 2 Answer the following questions regarding expandability of the current network and new network. (1) Fill in the appropriate site names for boxes through in the text. f h (2) Fill in the appropriate words or phrases for box in the text. i Sub-Question 3 Answer the following questions regarding the server used for the integrated business AP and the current network. (1) Explain in 40 words or less the problem which was resolved by abolishing the servers used for the distributed business AP and installing a server for the integrated business AP in the main office. (2) Explain in 60 words or less why an increase in traffic between the plant and the main office adversely affected other sites. Be sure to explain this from a network configuration perspective including the names of sites affected. Sub-Question 4 Answer the following questions regarding the connection test between the regional sales offices and the main office. (1) Fill in the appropriate words or phrases for box in the text. j (2) What is the port no. sent by the PORT command in line in Figure 4. (3) ISDN routers not only convert the IP header, but the IP address as well. Explain in 100 words or less specifically what is converted and how it is converted. (4) Explain in 60 words or less why communications were disconnected when the ISDN router was set as shown in the table. (5) Explain in 20 words or less which information of the packet monitoring data was useful in troubleshooting when the IPsec function was not used. Sub-Question 5 Answer the following questions regarding the new network. (1) Although the current network has some reliability problems concerning communications excluding the regional sales offices, these problems will be resolved by converting to the new network. Explain in 80 words or less the problems that will be resolved from a network configuration perspective. (2) If the IPsec function is implemented using PCs at the regional sales offices as Mr. T proposed, communication with the main office using existing ISDN routers would be impossible. Explain the reason for this in 90 words or less. (3) Explain in 100 words or less the specific features of and reasons for the IP address plan obtained using IP-VPN service when converting from the current network to the new network. 13
Q2. Read the following description of safety measures for a system connected to the Internet and answer Sub-Questions 1 through 5. Company Y is a large office supply sales company. It has a main office in Tokyo and 10 sales offices nationwide. With the rapid popularity of the Internet, one year ago Company Y began trial operation of a Web server for an Internet mail order business (hereafter referred to as "net business") system for contractant corporations. The total sales by net business increased suddenly, and at present has reached a level of 10 billion yen per month. As a result, Company Y has determined that direct net business is a promising business opportunity, and has decided to begin serious consideration of it by creating a special organization. Manager H who has been assigned to the net business has started working on the creation of a business plan. With costs of using the Internet decreasing and the population of Internet users increasing, the obstacles to using the Internet even among Company Y’s customer layer are decreasing and the expectation is that a usage environment will quickly be promoted. Since Company Y also thinks that there is plenty of opportunity to gain new customers due to the convenience of net business and merits associated with pricing, a business plan aiming at increasing the sales 10 times the current levels in three years is being created. Manager H has asked the Information Systems Department to investigate the systems which can handle electronic commerce on a scale of 100 billion yen per month based on the business plan created. Section Manager G of the Information Systems Department has asked Mr. F, who is in charge of Internet infrastructure, to do the study. Figure 1 shows the current configuration of Company Y’s LAN. ISP Router DMZ Web Mail DNS Corporate Firewall server server server Web server Public servers Internal segment DB Router Router PC PC server EC system DB server: Database server PC PC PC PC Fig. 1 Configuration of the LAN in the HQ 14
The EC system consists of a Web server and database server. The Web server of the EC system is located in the DMZ just like servers Company Y makes public to other companies such as the e-mail (hereafter referred to as “mail”) server, the DNS server, and the corporate Web server. Data such as customer data, order data and product data are stored in the database server. Since this was a trial system, no measures to cope with load increase or failure occurrences were included. In the last six months, Company Y has experienced the following four problems regarding the Internet. (1) Internet connections were lost due to the ISP. (2) Service stopped due to a failure of the Web server of the EC system. (3) Incidents occurred which caused stoppage of mail service and loss of mail. (4) There was damage due to a virus that infected the system through mail. As safety measures such as higher system reliability and performance to implement countermeasures against problems (1) through (4) at the same time, Mr. F put together an improvement plan with the following four points as essential features and reported it to Section Manager G. [Proposed System Improvements] (1) Connect with two ISPs (hereafter duplexing ISPs’). (2) Implement safety measures for main servers such as the EC system server, mail server, etc. (3) Adopt a configuration for the EC system that can handle sharp load increases. (4) Perform a virus check on mail. Section Manager G who was given an explanation of the proposed improvements. Specific understand the necessity of Mr. F’s proposal, and asked Mr. F to come up with a specific configuration design. Mr. F explained the proposed system improvements to Mr. B, of the system integration company to which the system construction was to be commissioned, and asked for a specific proposal. [Mr. B’s Proposal] (1) Duplexing ISPs Mr. B: There are two ways to duplex ISPs. The first is to operate an autonomous system, and the second is to connect to two ISPs using static routing just as has been done so far. 15
However, since the operation of the first type of system is difficult from a technical standpoint, and may have to take social responsibilities, it had best be avoided. Mr. F: Well, then, tell me about the second method. Mr. B: In this method, used to send and receive packets is divided for each public a server. The distribution route for packets transferred from the Internet is determined based on the IP address of the public server being accessed. Packets sent from to the b ISP can be allocated by a load balancing device. The load balancing device controls the assigned destination based on information in the transport layer and lower such as the IP address of the sender. Load balancing processing is performed based on a balancing algorithm that assigns destinations evenly. The load balancing device performs a periodic response check using “ping” and can switch transfer destinations if problems at the original transfer destination are detected. As long as the load balancing device has this function, it is possible to duplex ISPs. Fig. 2 shows the system configuration in this method. Mr. F: Please explain specifically how packets from the ISP are transferred in Fig. 2. Mr. B: IP addresses which are public in DMZ-1 are assumed to be borrowed from , and c IP addresses which are public in DMZ-2 are assumed to be borrowed from . d When configured this way, it is possible to distribute traffic between two ISPs because outside access to public servers in the DMZ is fixed to a path that is either ISP-1 or ISP-2 for each server. ISP-1 ISP-2 Company Y Router 1 Router 2 EC system Mail DNS Corporate Web Firewall 1 Firewall 2 Web server server server server DMZ-1 DMZ-2 LB Everything below this point omitted Fig. 2 System Configuration with Duplexed ISPs 16
Mr. F: How are packets from inside the company transferred to an ISP? Mr. B: Depending on the setting of the load balancing device, it is possible to evenly distribute packets between ISP-1 and ISP-2, or to usually transfer to ISP-1, but transfer to ISP-2 only when a problem with the route to ISP-1 occurs. If NAT is performed on the firewall at this time, (a) reply packets are returned to the same firewall through the same ISP. Mr. F: I understand. The method shown in Fig. 2 looks good. (2) Safety measures for the EC system and DNS server Mr. B: In order to handle future increases in load, a three-layer configuration should be used for the EC system with the addition of an application server (hereafter referred to as “AP servers”. With two Web servers for performing session management, screen control and task selection, expandability and flexibility can be increased by using the load balancing function of the load balancing device. With a configuration consisting of two application servers for executing application processing, load balancing is performed by using distribution server. The distribution server detects the CPU load ratio based on CPU performance and operational status, and possesses a function which select the server having the lowest load to execute processing. The database server has a cluster ring configuration. A RAID0+1 configuration is used for the disk device to increase speed and flexibility. RAID0 uses a method to spread data across several magnetic disks, while RAID1 is a method to perform for magnetic disk devices. RAID0+1 refers to a e method which performs both of these. Mr. F: What kinds of measures have you thought of for the other servers? Mr. B: The DNS server is split into two machines for internal and external use. (b) A DNS server for external use should be installed for each ISP out of consideration of possible ISP’s failure. I think there should be two DNS servers, the primary and secondary for internal use. (3) Safety measures for the mail server and virus protection Mr. F: What kind of safety measures have you thought of for the mail server? Mr. B: The mail system will consist of more than one server for safety. Server duplexing will be possible by setting mail distribution definitions on the server, and by using MX record setting information on the DNS server. The system will also check for viruses and clean any that are found. A detailed configuration is shown in Fig. 3. 17
ISP-1 ISP-2 Router 1 Router 2 SMTP server Anti-virus Spool server used for relay server SMTP server Anti-virus SMTP/POP used for relay server server LB PC Transfer route for mail received during normal operations Transfer route for mail received during system failure Transfer route for mail sent during normal operations Transfer route for mail sent during system failure Fig. 3 Duplexed Configuration for the Mail System Mr. F: How is mail from outside the company transferred? Mr. B: Mail from outside the company is first transferred to an SMTP server used for relay. The host names of the two SMTP servers used for relay can be registered by changing value for the DNS server for external use. When this is done, mail is usually a transferred via ISP-1, but if ISP-1 or the SMTP server used for relay fails, mail can be transferred to the other SMTP server used for relay via ISP-2. The SMTP servers used for relay are set so that received mail can be transferred to according to mail b distribution definitions. Backup server settings are included in the mail distribution definitions. is set so that mail is transferred to by accessing the DNS b c server for internal use. The host names for the two servers and are c d registered on the DNS server for internal use as MX records. By giving each of these a different value, backup can be performed if fails. Furthermore, the a c system can be set so that mail sent from a PC can be sent to the virtual IP address set by the load balancing device. The load balancing device can be set so that mail to the virtual IP address can be transferred along the transfer route used during normal operations. When mail is sent from a PC, the load balancing device can be used to perform failure recovery on the anti-virus server. 18
Mr. F: Is it possible to prevent the illegal relay of mail? Mr. B: Since mail transferred from inside the company always passes through an SMTP server used for relay, it is easy to take measures against illegal relay by the SMTP servers used for relay. Mr. F: That seems enough. It makes me rest easy. (4) Overall System Configuration Mr. B: Fig. 4 shows the overall configuration into which all of these are brought together. Safety measures are implemented for main servers according to the requirements. Network hardware such as the load balancing device, switching hub, and routers; the assignment server; and the firewall can all be exchanged relatively quickly. The duplex configuration is called a “cold standby” system. Mr. F: I understand. I will propose this configuration to my section manager. ISP-1 ISP-2 Router 1 Router 2 DNS DNS Corporate SMTP server SMTP server Firewall 1 Firewall 2 server 1 Web server server 2 used for relay used for relay DMZ-2 DMZ-1 LB LB Anti-virus server Web Web Distribution LB Switching server server server hub Anti-virus Application server Spool server server SMTP/POP Application Everything below server server this point omitted Disk Database DNS server for device server internal use Database Disk DNS server for server device internal use EC system Fig. 4 Configuration of the New LAN at HQ When Mr. F proposed the safety measures for the Internet-related system to Section Manager G based on Mr. B’s proposal, he was instructed to investigate the following five issues concerning the operation of the EC system. 19
[Issues concerning the operation of the EC system] Measures to prevent the operations from stopping even when ISP-2 fails. Measures to prevent the operations from stopping even during a power outage or when power supply to electronic hardware is cut off during periodic inspection. The operational organization needed for running the system 24 hours per day, 365 days per year. Measures for flexibly handling rapid increases in traffic. The operational organization needed for making backups in the middle of the night. Since it was difficult to resolve these issues using internal system operation, Mr. F thought that it would be a good idea to use the housing service of an Internet Data Center (IDC). So Mr. F decided to consult with Mr. C, a representative of an IDC, on the possibility of using housing services for the EC system. [Housing Services of the IDC] Mr. F: Although I am considering the system shown in Fig. 4 in terms of safety measures for an Internet system, there are several issues regarding operation, and so I’m investigating the possibility of using housing services. Maintenance of the EC system database and application programs, and the operation of services such as processing order data will be handled in-house as usual. However, I cannot create an organization that can work at midnight and on holidays. Can this problem be solved by using housing services? Mr. C: If you choose the housing of the entire EC system and hardware related to DMZ-2 (hereafter referred to as “Proposal 1”), you can expect to resolve all problems. If you choose the housing of only hardware related DMZ-2 (hereafter referred to as “Proposal 2”), several problems will remain. Mr. F: But in the case of Proposal 1, it’s necessary to change the system configuration shown in Fig. 4, right? Mr. C: Yes, that’s right. the slight configuration changes shown in Fig. 5 are needed. 20