7.7 Summary

We have seen that there is a confusing arsenal of tools available to assist in improving internetwork performance—some standard and some proprietary. Many of these solutions can be used together to good effect, although you should exercise caution and ensure that the solutions used are complementary. It is especially important as part of any optimization process to first baseline the network, identify potential bottlenecks, and only then begin to change operating parameters.

In this chapter, we also discussed the following:

Routing and addressing hierarchy is key for large-scale internetwork design. It is critical that router performance is optimized and that valuable WAN bandwidth is preserved by using features such as route summarization where possible.
Load splitting and clustering techniques are often the only effective way to provide scalability for CPU-intensive devices such as firewalls and Web servers.
Caching is a generic technique that is used in many areas of IT. The principal theme of caching is to localize copies of information that are used frequently, in order to minimize the number of remote transactions and reduce latency. Caching has become especially important as a way of improving performance for Web-based applications.
Bandwidth managers and load balancers play complementary roles. Proxy servers such as Squid work with any of these technologies.
There could be potential conflicts between standalone bandwidth managers and routers that implement their own prioritization and queuing techniques. Some bandwidth managers, for example, claim that by shaping traffic they prevent router prioritization routines from ever being invoked in the first place, thus avoiding any conflicts.
Data compression is a good generic technique for use on low-bandwidth wide area links (typically below 2 Mbps). It may be implemented in hardware or software. It is also applicable for optimizing storage media such as hard disks.
Storage is becoming a critical resource for large-scale enterprise networks. Techniques such as Hierarchical Storage Management (HSM), Network Attached Storage (NAS), Storage Attached Networks (SANs), and Optical Storage Networks (OSNs) are increasingly being deployed to solve the space and performance problems associated with managing and moving terabytes of data over today's networks.

References

[1] T. Kenyon, High-Performance Network Design: Design Techniques and Tools (Woburn, MA: Digital Press, 2001).

[2] R. Jain, The Art of Computer Systems Performance Analysis (New York: John Wiley & Sons, 1991).

[3] L. Chappell, Novell's Guide to NetWare LAN Analysis (Alameda, CA: Sybex, Inc., 1993).

[4] G. S. Sidhu, R. F. Andrews, and A. B. Oppenheimer, Inside Apple-Talk (Reading, MA: Addison-Wesley, 1993).

[5] M. A. Weiss, Data Structures and Algorithm Analysis in C, 2nd ed. (Reading, MA: Addison-Wesley, 1997).

[6] A. Lempel and J. Ziv, "A Universal Algorithm for Sequential Data Compression," IEEE Transactions on Information Theory, vol. IT-23, No. 3, May 1977.

[7] PPP Stac LZS Compression Protocol, RFC 1974, August 1996.

[8] www.itu.ch, The International Telecommunications Union (ITU) home page.

[9] Compressing TCP/IP Headers for Low-Speed Serial Links, RFC 1144.

[10] The PPP Compression Control Protocol (CCP), RFC 1962, July 1996.

[11] D. Saloman, Data Compression The Complete Reference (New York: Springer-Verlag, 1998).

[12] L. Kleinrock, Queuing Systems, vol. 2: Computer Applications (New York: Wiley Interscience, 1975).

[13] S. Keshav, An Engineering Approach to Computer Networking (Reading, MA: Addison-Wesley, 1997).

[14] ftp.ee.lbl.gov/floyd/cbq.html, The CBQ Web page—useful links to papers and source code for CBQ.

[15] ftp://ftp.ee.lbl.gov/papers/admit.ps.Z, Comments on measurement-based admissions control for controlled-load services.

[16] ftp.ee.lbl.gov/floyd/cbq.html, Notes on CBQ and guaranteed service.

[17] R. Jain and K. K. Ramakrishnan, "Congestion Avoidance in Computer Networks with a Connectionless Network Layer: Concepts, Goals, and Methodology," in Proceedings of the IEEE Computer Networking Symposium, Washington, DC, April 1988.

[18] K. K. Ramakrishnan and R. Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks," ACM Transactions on Computer Systems, vol. 8, no. 2, 1990.

[19] S. Floyd and V. Jacobson, "Random Early Detection Gateways for Congestion Avoidance," IEEE/ACM Transactions on Networking, August 1993.

[20] S. Floyd and K. Fall, "Router Mechanisms to Support End-to-End Congestion Control," in Proceedings of SIGCOMM '97, 1997.

[21] www-nrg.ee.lbl.gov/floyd/, Sally Floyd home page; excellent links for research papers on congestion avoidance and queuing strategies.

[22] P. Albitz, and C. Liu, DNS and BIND (Cambridge MA: O'Reilly & Associates, 1997).

[23] www.isc.org, The Internet Software Consortium Web site—contains source for the BIND application.

[24] Domain Names—Concepts and Facilities, RFC 1034, November 1987.

[25] Domain Names—Implementation and Specification, RFC 1035, November 1987.

[26] DNS Extensions to Support IP Version 6, RFC 786, December 1995.

[27] DNS Support for Load Balancing, RFC 1794, April 1995.

[28] Ethernet Address Resolution Protocol: Converting Network Protocol Addresses to 48.bit Ethernet Addresses for Transmission on Ethernet Hardware, RFC 826, November 1982.

[29] The IP Network Address Translator (NAT), RFC 1631, May 1994.

[30] Load Sharing Using IP Network Address Translation (LSNAT), RFC2391, August 1998.

[31] Internet Cache Protocol (ICP), Version 2, RFC 276, September 1997.

[32] Application of Internet Cache Protocol (ICP), Version 2, RFC 277, September 1997.

[33] Network News Transfer Protocol, RFC 977, February 1986.

[34] Cisco Web Cache Control Protocol V1.0. Internet Draft draft-ietf-wrec-web-pro-00.txt, June 1999.

[35] Hyper Text Caching Protocol (HTCP/0.0), RFC 2756, January 2000. (Experimental).

[36] D. A. Menasce and A. F. Almeida, Capacity Planning for Web Performance (Englewood Cliffs, NJ: Prentice Hall, 1998).

[37] www.cacheflow.com, Caheflow home Web site—useful white papers on caching technology.

[38] ftp://ircache.nlanr.net/Traces/8, Weekly access log files maintained by the National Laboratory for Applied Network Research (NLANR).

[39] Benchmarking Methodology for Network Interconnect Devices, RFC 2544, March 1999.

[40] A. Orda and R. Rom, "Optimal Packet Fragmentation in Computer Networks," Department of Electrical Engineering, Technion—Israel Institute of Technology, Haifa, Israel, March 1994.

[41] Congestion Control in IP/TCP Internetworks, RFC 896, January 1984.

[42] Path MTU Discovery, RFC 1191, November 1990.

[43] Internet Control Message Protocol, RFC 792, September 1981.

[44] TCP Extensions for High Performance, RFC 1323, May 1992.

[45] www.ibm.com, IBM home Web site.