Managing Bandwidth: Deploying Qos in Enterprise Networks

by Croll, Alistair; Packman, Eric

• ISBN13: 9780130113917
• ISBN10: 0130113913
• Edition: 1st
• Format: Hardcover
• Copyright: 1999-07-01
• Publisher: PRENTICE HALL

Summary

To build an enterprise network that delivers real value to the business, it's no longer enough to simply add bandwidth: you must manage your bandwidth more effectively than ever before. This book shows you how, helping you reduce costs, delay expenditures, and deliver new applications with precisely service quality they require. Understand the technologies and business trends that are driving service level management in the enterprise network. Learn advanced techniques for differentiating between low-priority and high-priority applications; then delivering bandwidth in the appropriate quantities, within appropriate latency and jitter parameters. Compare Class of Service (CoS) approaches with Quality of Service (QoS) approaches such as ATM's QoS and Resource Reservation Protocol for IP networks. Understand how to establish "A" list customers and give them preferred access to Internet and other network resources; handle peak loads more effectively; delay network upgrades; and much more. The book includes four detailed case studies representing financial services, consulting, retail and academic organizations. For every networking professional responsible for maximizing the value and effectiveness of enterprise networks.

Table of Contents

Introduction	xiii
Intended audience	xv	(1)
Acknowledgments	xvi	(1)
Recent Developments	xvi
Part 1	1	(62)
1 Defining the scope of bandwidth management	3	(18)
Characteristics of a bandwidth-managed network	4	(6)
Rolling your own: How close to the edge do you want to live?	10	(1)
Alphabet soup: QOS, COS, integrated and differentiated services	11	(2)
How real is this?	13	(1)
Comparing reservation, network COS, and media COS	14	(7)
2 Motivations for bandwidth management	21	(20)
Do we need true QOS end-to-end?	23	(2)
Is bandwidth management necessary?	25	(1)
The argument for big bandwidth	25	(2)
The argument for managed bandwidth	27	(6)
Technologies that enable multiservice networks	33	(8)
3 A history of networking usage	41	(14)
How things change	42	(1)
Early master-slave computer systems	42	(2)
The arrival of the LAN	44	(1)
The mission-critical business network	45	(1)
Distributing centralized applications	46	(1)
The wrong way: The LAN as a hard drive	46	(1)
Client-server models	47	(2)
The managed-chaos business network	49	(1)
Peer-to-peer networks	50	(2)
Manager-agent networks	52	(3)
4 Bandwidth today	55	(10)
The next-generation business network	56	(1)
The WAN and bandwidth	57	(1)
Virtual centralization	58	(1)
Who'll go first?	59	(1)
Other factors	60	(3)
Part 2	63	(50)
5 Understanding traffic performance characteristics	65	(26)
Sources of delay	67	(3)
Traffic performance characteristics2	70	(4)
Retransmission	74	(3)
Capacity	77	(5)
Delay (latency)	82	(9)
6 The needs of different traffic types	91	(24)
Voice	93	(9)
Video	102	(4)
Interactive conversation	106	(1)
The converged network	106	(4)
Obstacles to convergence	110	(3)
Part 3	113	(208)
7 An introduction to traffic management	115	(14)
A policy system	117	(3)
A high-level view: Classify, handle, police, and monitor	120	(5)
Stepping back from QOS deployment: A strategic perspective	125	(4)
8 Classification--identifying the traffic	129	(56)
Where on the network does classification make sense?	130	(8)
Classification down the stack	138	(2)
The highest level: User identification	140	(5)
Top-of-the-stack classification: Application	145	(1)
Between flow and application: OSI leftovers	146	(1)
Transport-level classification: Flow	147	(2)
Network prioritization classification: IP TOS	149	(11)
Differentiated Services work and the DS field	160	(8)
Layer-3 address classification	168	(1)
Layer-2 address classification: MAC information	169	(1)
Link-prioritization classification: 802.1Q/p	169	(11)
Link prioritization classification: VLAN membership	180	(2)
Link prioritization classification: ATM QOS	182	(3)
9 Complications to classification	185	(12)
Network address translation	186	(3)
Stateful traffic	189	(3)
VPN traffic	192	(1)
Conditional statements	193	(1)
Extrinsic conditions	194	(1)
Congestion conditions	195	(2)
10 Traffic handling within a device	197	(34)
Common ground	198	(2)
Queues, buckets, and admission control	200	(2)
Queuing systems and prioritization	202	(6)
Discarding traffic	208	(8)
Rate controllers	216	(1)
Bandwidth reservation	217	(4)
RSVP	221	(7)
Putting it all together	228	(3)
11 Traffic handling by path selection	231	(26)
Marking the traffic: Ethernet prioritization	232	(5)
Marking the traffic: Leveraging your ISP	237	(2)
Forwarding to a permanent circuit	239	(3)
Dynamic circuit creation	242	(15)
12 Server-side delay	257	(18)
Session load balancing	258	(4)
Service load balancing	262	(2)
Caching	264	(7)
Co-processing	271	(1)
The IP front-end processor	271	(1)
A summary of handling techniques	272	(3)
13 Directory infrastructures	275	(12)
Directory deployment today	277	(1)
Basic components of a directory service	278	(3)
The X.500 directory standard	281	(1)
Common data structures	282	(1)
Deciding to agree: The DEN ad-hoc working group	283	(1)
LDAP	283	(2)
Putting directories to work	285	(2)
14 Policy systems	287	(22)
Policy protocols	288	(3)
Radius	291	(1)
Diameter	292	(1)
COPS	293	(8)
Publishing	301	(1)
Policy across administrative domains	302	(3)
Publishing policies	305	(4)
15 Monitoring service levels in a network	309	(12)
Service-level agreements	310	(4)
Metrics	314	(4)
Monitoring the condition of the network with traditional tools	318	(3)
Part 4	321	(50)
16 The shape of bandwidth-managed networks	323	(10)
Traditional networking: L2 on the edge, L3 at the core	324	(1)
Next-generation networks: L2 at the core and edge	325	(3)
Real-time, tactical, and strategic	328	(1)
Real-time measures	329	(2)
Operational measures	331	(1)
Strategic measures	332	(1)
17 Practical bandwidth management	333	(10)
Where to begin?	334	(1)
Ready for real time?	334	(1)
Deployment roadmap	335	(3)
Building a differentiated-services strategy	338	(1)
Picking a philosophy	339	(1)
Putting it together: Multiple mapping systems	340	(3)
18 Case studies	343	(28)
Case study: A distributed retail organization	345	(4)
Case study: An academic campus	349	(4)
Case study: A centralized financial institution	353	(4)
Case study: A 2-coast consulting firm	357	(6)
Case study: A distributed manufacturing company	363	(3)
Case study: A small business	366	(5)
Part 5	371	(14)
19 Conclusions	373	(12)
Consequences for IT managers	374	(2)
Consequences for designers of network services	376	(3)
Consequences for those who run networks	379	(1)
Consequences for application developers	380	(3)
A final word	383	(2)
Part 6	385	(28)
A An overview of TCP	387	(10)
TCP setup	388	(7)
When is UDP better than TCP?	395	(2)
B Glossary	397	(10)
C Bibliography and references	407	(6)
Web sites	408	(1)
Requests for Comments (RFCs) and Internet Drafts	409	(4)
Index	413