15 440 InterDomain Routing BGP Border Gateway Protocol
15 -440 Inter-Domain Routing BGP (Border Gateway Protocol) DNS (Domain Name System) These slides proudly ripped from Srini Seshan and Dave Anderson and Seth Goldstein, 15 -441 F’ 06 and F’ 08
Outline • Internet Structure/Routing Hierarchy • External BGP (E-BGP) • Internal BGP (I-BGP)
A Logical View of the Internet? • After looking at RIP/OSPF descriptions • End-hosts connected to routers • Routers exchange messages to determine connectivity • NOT TRUE! R R R
Internet’s Area Hierarchy • What is an Autonomous System (AS)? • A set of routers under a single technical administration, using an interior gateway protocol (IGP) and common metrics to route packets within the AS and using an exterior gateway protocol (EGP) to route packets to other AS’s • Each AS assigned unique ID • AS’s peer at network exchanges
AS Numbers (ASNs) ASNs are 16 bit values 64512 through 65535 are “private” Currently over 15, 000 in use • • Genuity: 1 MIT: 3 CMU: 9 UC San Diego: 7377 AT&T: 7018, 6341, 5074, … UUNET: 701, 702, 284, 12199, … Sprint: 1239, 1240, 6211, 6242, … … ASNs represent units of routing policy
Example 1 2 IGP 2. 1 IGP EGP 1. 1 2. 2. 1 1. 2 EGP EGP 3 IGP 5 EGP 3. 1 5. 2 2. 2 IGP 3. 2 4. 1 IGP 4. 2 4
A Logical View of the Internet? • RIP/OSPF not very scalable area hierarchies • NOT TRUE EITHER! • ISP’s aren’t equal • Size • Connectivity ISP R R R
A Logical View of the Internet • Tier 1 ISP • “Default-free” with global reachability info • Tier 2 ISP Tier 3 Tier 2 • Regional or country-wide • Tier 3 ISP Tier 2 • Local Customer Provider Tier 1 Tier 2
Transit vs. Peering Transit ($$ 1/2) Transit ($$$) ISP Y ISP P Transit ($) Transit ($$$) ISP Z Transit ($$) Transit ($$$) Peering Transit ($$) ISP X Transit ($$)
Policy Impact • “Valley-free” routing • Number links as (+1, 0, -1) for provider, peer and customer • In any path should only see sequence of +1, followed by at most one 0, followed by sequence of -1 • WHY? • Consider the economics of the situation
Outline • Internet Structure/Routing Hierarchy • External BGP (E-BGP) • Internal BGP (I-BGP)
Choices • Link state or distance vector? • No universal metric – policy decisions • Problems with distance-vector: • Bellman-Ford algorithm may not converge • Problems with link state: • Metric used by routers not the same – loops • LS database too large – entire Internet • May expose policies to other AS’s
Solution: Distance Vector with Path • Each routing update carries the entire path • Loops are detected as follows: • When AS gets route, check if AS already in path • If yes, reject route • If no, add self and (possibly) advertise route further • Advantage: • Metrics are local - AS chooses path, protocol ensures no loops
Interconnecting BGP Peers • BGP uses TCP to connect peers • Advantages: • Simplifies BGP • No need for periodic refresh - routes are valid until withdrawn, or the connection is lost • Incremental updates • Disadvantages • Congestion control on a routing protocol? • Poor interaction during high load
Hop-by-hop Model • BGP advertises to neighbors only those routes that it uses • Consistent with the hop-by-hop Internet paradigm • e. g. , AS 1 cannot tell AS 2 to route to other AS’s in a manner different than what AS 2 has chosen (need source routing for that) • BGP enforces policies by choosing paths from multiple alternatives and controlling advertisement to other AS’s
Examples of BGP Policies • A multi-homed AS refuses to act as transit • Limit path advertisement • A multi-homed AS can become transit for some AS’s • Only advertise paths to some AS’s • An AS can favor or disfavor certain AS’s for traffic transit from itself
BGP Messages • Open • Announces AS ID • Determines hold timer – interval between keep_alive or update messages, zero interval implies no keep_alive • Keep_alive • Sent periodically (but before hold timer expires) to peers to ensure connectivity. • Sent in place of an UPDATE message • Notification • Used for error notification • TCP connection is closed immediately after notification
BGP UPDATE Message • List of withdrawn routes • Network layer reachability information • List of reachable prefixes • Path attributes • Origin • Path • Metrics • All prefixes advertised in message have same path attributes
Path Selection Criteria • Attributes + external (policy) information • Examples: • Hop count • Policy considerations • Preference for AS • Presence or absence of certain AS • Path origin • Link dynamics
LOCAL PREF • Local (within an AS) mechanism to provide relative priority among BGP routers (e. g. R 3 over R 4) R 5 R 1 R 2 AS 200 AS 100 R 3 Local Pref = 500 AS 256 AS 300 Local Pref = 800 I-BGP R 4
LOCAL PREF – Common Uses • Peering vs. transit • Prefer to use peering connection, why? • In general, customer > peer > provider • Use LOCAL PREF to ensure this
AS_PATH • List of traversed AS’s AS 200 170. 10. 0. 0/16 AS 100 180. 10. 0. 0/16 AS 300 AS 500 180. 10. 0. 0/16 300 200 170. 10. 0. 0/16 300 200
Multi-Exit Discriminator (MED) • Hint to external neighbors about the preferred path into an AS • Non-transitive attribute • Different AS choose different scales • Used when two AS’s connect to each other in more than one place
MED • Hint to R 1 to use R 3 over R 4 link • Cannot compare AS 40’s values to AS 30’s 180. 10. 0. 0 MED = 50 R 1 AS 10 R 3 AS 30 180. 10. 0. 0 MED = 120 R 2 AS 40 180. 10. 0. 0 MED = 200 R 4
MED • MED is typically used in provider/subscriber scenarios • It can lead to unfairness if used between ISP because it may force one ISP to carry more traffic: SF ISP 1 ISP 2 • ISP 1 ignores MED from ISP 2 • ISP 2 obeys MED from ISP 1 • ISP 2 ends up carrying traffic most of the way NY
Decision Process • Processing order of attributes: • Select route with highest LOCAL-PREF • Select route with shortest AS-PATH • Apply MED (if routes learned from same neighbor)
Important Concepts • Wide area Internet structure and routing driven by economic considerations • Customer, providers and peers • BGP designed to: • Provide hierarchy that allows scalability • Allow enforcement of policies related to structure • Mechanisms • Path vector – scalable, hides structure from neighbors, detects loops quickly
What is DNS? • DNS (Domain Name Service) is primarily used to translate human readable names into machine usable addresses, e. g. , IP addresses. • DNS goal: • Efficiently locate resources. E. g. , Map name IP address • Scale to many users over a large area • Scale to many updates Lecture 13 15 -441 © 2008 36
How resolve name IP addr? Lecture 13 15 -441 © 2008 37
Obvious Solutions (1) Why not centralize DNS? • Single point of failure • Traffic volume • Distant centralized database • Single point of update • Doesn’t scale! Lecture 13 15 -441 © 2008 38
Obvious Solutions (2) Why not use /etc/hosts? • Original Name to Address Mapping • • Flat namespace /etc/hosts SRI kept main copy Downloaded regularly • Mid 80’s this became untenable. Why? • Count of hosts was increasing: machine per domain machine per user • Many more downloads • Many more updates Lecture 13 15 -441 © 2008 /etc/hosts still exists. 39
Domain Name System Goals • Basically a wide-area distributed database (The biggest in the world!) • Scalability • Decentralized maintenance • Robustness • Global scope • Names mean the same thing everywhere • Don’t need all of ACID • Atomicity • Strong consistency • Do need: distributed update/query & Performance Lecture 13 15 -441 © 2008 40
Programmer’s View of DNS • Conceptually, programmers can view the DNS database as a collection of millions of host entry structures: /* DNS host entry structure */ struct hostent { char *h_name; /* official domain name of host */ char **h_aliases; /* null-terminated array of domain names */ int h_addrtype; /* host address type (AF_INET) */ • in_addr is a struct consisting of 4 -bytein. IPbytes addr*/ int h_length; /* length of an address, **h_addr_list; /* null-termed array offrom in_addr structs */ • char Functions for retrieving host entries DNS: }; • gethostbyname: query key is a DNS host name. • gethostbyaddr: query key is an IP address. Lecture 13 15 -441 © 2008 41
DNS Message Format 12 bytes Identification Flags No. of Questions No. of Answer RRs No. of Authority RRs No. of Additional RRs Name, type fields for a query Questions (variable number of answers) RRs in response to query Answers (variable number of resource records) Records for authoritative servers Additional “helpful info that may be used Lecture 13 Authority (variable number of resource records) Additional Info (variable number of resource records) 15 -441 © 2008 42
DNS Header Fields • Identification • Used to match up request/response • Flags • • Lecture 13 1 -bit to mark query or response 1 -bit to mark authoritative or not 1 -bit to request recursive resolution 1 -bit to indicate support for recursive resolution 15 -441 © 2008 43
DNS Records RR format: (class, name, value, type, ttl) • DB contains tuples called resource records (RRs) • Classes = Internet (IN), Chaosnet (CH), etc. • Each class defines value associated with type For “IN” class: • Type=CNAME • Type=A • name is hostname • value is IP address • Type=NS • name is an alias name for some “canonical” name • value is canonical name • name is domain (e. g. foo. com) • Type=MX • value is hostname of • value is name of authoritative mailserver associated with name server for this domain name Lecture 13 15 -441 © 2008 44
Properties of DNS Host Entries Different kinds of mappings are possible: • 1 -1 mapping between domain name and IP addr: provolone. crcl. cs. cmu. edu maps to 128. 2. 218. 81 • Multiple domain names maps to the same IP addr: www. scs. cmu. edu and www. cs. cmu. edu both map to 128. 2. 203. 164 • Single domain name maps to multiple IP addresses: aol. com and www. aol. com map to multiple IP addrs. • Some valid domain names don’t map to any IP addr: crcl. cs. cmu. edu doesn’t have a host Lecture 13 15 -441 © 2008 45
DNS Design: Hierarchy Definitions root org net gwu ucb edu com uk cmu bu mit cs ece crcl Lecture 13 • Each node in hierarchy stores a list of names that end with same suffix • Suffix = path up tree • E. g. , given this tree, where would following be stored: • Fred. com • Fred. edu • Fred. cmu. edu • Fred. crcl. cs. cmu. edu • Fred. cs. mit. edu 15 -441 © 2008 46
DNS Design: Zone Definitions root org net gwu ucb edu com uk bu mit cmu cs ece crcl • Zone = contiguous section of name space • E. g. , Complete tree, single node or subtree • A zone has an associated set of name servers • Must store list of names and tree links Subtree Single node Complete Tree Lecture 13 15 -441 © 2008 47
DNS Design: Cont. • Zones are created by convincing owner node to create/delegate a subzone • Records within zone stored in multiple redundant name servers • Primary/master name server updated manually • Secondary/redundant servers updated by zone transfer of name space • Zone transfer is a bulk transfer of the “configuration” of a DNS server – uses TCP to ensure reliability • Example: • CS. CMU. EDU created by CMU. EDU admins • Who creates CMU. EDU or. EDU? Lecture 13 15 -441 © 2008 48
DNS: Root Name Servers • Responsible for “root” zone • 13 root name servers • Currently {a-m}. root-servers. net • Local name servers contact root servers when they cannot resolve a name • Why 13? Lecture 13 15 -441 © 2008 49
Not really 13! Check out anycast) Lecture 13 10/08, from 15 -441 © 2008 www. root-servers. org 50
So Far • Database structure • Hierarchy of labels x. y. z • Organized into zones • Zones have nameservers (notice plural!) • Database layout • Records which map names, names ip, etc. • Programmer API: gethostbyname, … Lecture 13 15 -441 © 2008 51
Servers/Resolvers • Each host has a resolver • Typically a library that applications can link to • Local name servers hand-configured (or DHCP) (e. g. /etc/resolv. conf) • Name servers • Either responsible for some zone or… • Local servers • Do lookup of distant host names for local hosts • Typically answer queries about local zone Lecture 13 15 -441 © 2008 52
Typical Resolution edu. u m c. www. cs. cmu. edu s u c. d e w. u ww m c. 1 s n NS NS ns 1. cs. cmu. e du Local Client DNS server A ww w= IPa dd r root & edu DNS server ns 1. cmu. edu DNS server ns 1. cs. cmu. edu DNS server Hmm: Notice root server returned NS ns 1. cmu. edu Lecture 13 15 -441 © 2008 53
Typical Resolution • Steps for resolving www. cmu. edu • • • Application calls gethostbyname() (RESOLVER) Resolver contacts local name server (S 1) S 1 queries root server (S 2) for (www. cmu. edu) S 2 returns NS record for cmu. edu (S 3) What about A record for S 3? This is what the additional info section is for (PREFETCHING) • S 1 queries S 3 for www. cmu. edu • S 3 returns A record for www. cmu. edu • Can Lecture 13 return multiple 15 -441 A © records 2008 54
Lookup Methods Recursive query: root name server • Server goes out and searches for more info • Only returns final answer or “not found” 2 iterated query 3 Iterative query: 4 7 • Server responds with local name server intermediate name server as much as it knows. dns. eurecom. fr dns. umass. edu 5 6 authoritative name • “I don’t know this 1 8 server name, but ask this dns. cs. umass. edu server” requesting host surf. eurecom. fr Workload impact on choice? gaia. cs. umass. edu • Root/distant server does iterative • Local server typically does recursive Lecture 13 15 -441 © 2008 55
How to manage workload? • Does root nameserver do recursive lookups? • What about other zones? • What about imbalance in popularity? • . com versus. dj • google. com versus bleu. crcl. cs. cmu. edu? • How do we scale query workload? Lecture 13 15 -441 © 2008 56
Workload and Caching • DNS responses are cached • Quick response for repeated translations • Other queries may reuse some parts of lookup • E. g. , NS records for domains • DNS negative queries are cached • Don’t have to repeat past mistakes • E. g. , misspellings, search strings in resolv. conf • How do you handle updates? Lecture 13 15 -441 © 2008 57
Workload and Caching • DNS responses are cached • Quick response for repeated translations • Other queries may reuse some parts of lookup • E. g. , NS records for domains • DNS negative queries are cached • Don’t have to repeat past mistakes • E. g. , misspellings, search strings in resolv. conf • Cached data periodically times out • Lifetime (TTL) of data controlled by owner of data • TTL passed with every record Lecture 13 15 -441 © 2008 58
Typical Resolution edu. u m c. www. cs. cmu. edu s u c. d e w. u ww m c. 1 s n NS NS ns 1. cs. cmu. e du Local Client DNS server A ww w= IPa dd r Lecture 13 15 -441 © 2008 root & edu DNS server ns 1. cmu. edu DNS server ns 1. cs. cmu. edu DNS server 59
Subsequent Lookup Example root & edu DNS server ftp. cs. cmu. edu Client Local DNS server ftp . cs Af . cm tp= u. e IPa dd du r Lecture 13 15 -441 © 2008 cmu. edu DNS server cs. cmu. edu DNS server 60
Reliability • DNS servers are replicated • Name service available if ≥ one replica is up • Queries can be load balanced between replicas • UDP used for queries • Need reliability must implement this on top of UDP! • Why not just use TCP? • Try alternate servers on timeout • Exponential backoff when retrying same server • Same identifier for all queries • Don’t care which server responds Lecture 13 15 -441 © 2008 61
So far • Hierarchial name space Lecture 13 15 -441 © 2008 62
Reverse DNS Arpa: backronym Address and Routing Parameter Area • Task unnamed root edu arpa in-addr 128 2 cmu cs crcl 204 27 Lecture 13 bleu 128. 2. 204. 27 • Given IP address, find its name • Method • Maintain separate hierarchy based on IP names • Write 128. 2. 204. 27 as 27. 204. 2. 128. in-addr. arpa • Why is the address reversed? • Managing • Authority manages IP addresses assigned to it • E. g. , CMU manages name 15 -441 © 2008 63
. arpa Name Server Hierarchy in-addr. arpa a. root-servers. net • • • m. root-servers. ne chia. arin. net 128 (dill, henna, indigo, epazote, figwort, gin cucumber. srv. cs. cmu. edu, 2 t-ns 1. net. cmu. edu t-ns 2. net. cmu. edu mango. srv. cs. cmu. edu 204 (peach, banana, blueberry) bleu 128. 2. 204. 27 Lecture 13 • At each level of hierarchy, have group of servers that are authorized to handle that region of hierarchy 15 -441 © 2008 64
Prefetching • Name servers can additional data to response • Why would they? Lecture 13 15 -441 © 2008 65
Prefetching • Name servers can additional data to response • Why would they? • Typically used for prefetching • CNAME/MX/NS typically point to another host name • Responses include address of host referred to in “additional section” Lecture 13 15 -441 © 2008 66
Mail Addresses • MX records point to mail exchanger for a name • E. g. cmu. edu. 2590 IN IN MX MX 10 CMU-MX 4. ANDREW. cmu. edu. 10 CMU-MX 5. ANDREW. cmu. edu. • Addition of MX record type proved to be a challenge • How to get mail programs to lookup MX record for mail delivery? • Needed critical mass of such mailers • Could we add a new one now? Lecture 13 15 -441 © 2008 67
Outline • DNS Design • DNS Today Lecture 13 15 -441 © 2008 68
Root Zone • Generic Top Level Domains (g. TLD) =. com, . net, . org, etc… • Country Code Top Level Domain (cc. TLD) =. us, . ca, . fi, . uk, etc… • Root server ({a-m}. root-servers. net) also used to cover g. TLD domains • Load on root servers was growing quickly! • Moving. com, . net, . org off root servers was clearly necessary to reduce load done Aug 2000 • How significant an effect would this have? • On load? • On performance? Lecture 13 15 -441 © 2008 69
g. TLDs • Unsponsored • • . com, . edu, . gov, . mil, . net, . org. biz businesses. info general info. name individuals • Sponsored (controlled by a particular association) • • . aero air-transport industry. catalan related. coop business cooperatives Is there anything special about. com? . jobs job announcements about adding. goldstein as a g. TLD? . museum. What museums. pro accountants, lawyers, and physicians. travel industry • Starting up • . mobile phone targeted domains • . postal • . telephone related • Proposed • . asia, . cym, . geo, . kid, . mail, . sco, . web, . xxx • Whatever you want! Lecture 13 15 -441 © 2008 70
New Registrars • Network Solutions (NSI) used to handle all registrations, root servers, etc… • Clearly not the democratic (Internet) way • Large number of registrars that can create new domains However NSI still handles A root server Lecture 13 15 -441 © 2008 71
• Measurements of DNS No centralized caching per site • Each machine runs own caching local server • Why is this a problem? • How many hosts do we need to share cache? recent studies suggest 10 -20 hosts • “Hit rate for DNS: 1 - (#DNS/#connections) 80% • Is this good or bad? • Most Internet traffic was Web with HTTP 1. 0 • What does a typical page look like? average of 4 -5 imbedded objects needs 4 -5 transfers • This alone accounts for 80% hit rate! • Lower TTLs for A records does not affect performance • DNS performance really relies more on NS-record caching Lecture 13 15 -441 © 2008 72
• Measurements of DNS No centralized caching per site • Each machine runs own caching local server • Why is this a problem? • How many hosts do we need to share cache? recent studies suggest 10 -20 hosts • “Hit rate for DNS: 1 - (#DNS/#connections) 80% • Is this good or bad? • Most Internet traffic was Web with HTTP 1. 0 • What does a typical page look like? average of 4 -5 imbedded objects needs 4 -5 transfers • This alone accounts for 80% hit rate! • Lower TTLs for A records does not affect performance • DNS performance really relies more on NS-record caching Lecture 13 15 -441 © 2008 73
Tracing Hierarchy (1) • Dig Program • Allows querying of DNS system • Use flags to find name server (NS) • Disable recursion so that operates one step at a time unix> dig +norecurse @a. root-servers. net NS Zone kittyhawk. cmcl. cs. cmu. edu TTL ; ; AUTHORITY SECTION: Type IN edu. 172800 NS L 3. NSTLD. COM. Value edu. 172800 IN NS Class D 3. NSTLD. COM. edu. 172800 IN NS • All. edu names handled by set of servers A 3. NSTLD. COM. Lecture 13 15 -441 © 2008 74
Tracing Hierarchy (2) • 3 servers handle CMU names unix> dig +norecurse @e 3. nstld. com NS kittyhawk. cmcl. cs. cmu. edu ; ; AUTHORITY SECTION: cmu. edu. 172800 IN Lecture 13 NS NS NS 15 -441 © 2008 CUCUMBER. SRV. cs. cmu. edu. T-NS 1. NET. cmu. edu. T-NS 2. NET. cmu. edu. 75
Tracing Hierarchy (3 & 4) • 4 servers handle CMU CS names unix> dig +norecurse @t-ns 1. net. cmu. edu NS kittyhawk. cmcl. cs. cmu. edu ; ; AUTHORITY SECTION: cs. cmu. edu. 86400 IN NS MANGO. SRV. cs. cmu. edu. • Quasar is master NS for this cs. cmu. edu. 86400 IN zone NS PEACH. SRV. cs. cmu. edu. IN NS unix>dig +norecurse 86400 @blueberry. srv. cs. cmu. edu BANANA. SRV. cs. cmu. edu. NS kittyhawk. cmcl. cs. cmu. edu. 86400 IN NS ; ; BLUEBERRY. SRV. cs. cmu. edu. AUTHORITY SECTION: Lecture 13 15 -441 © 2008 76
DNS (Summary) • Motivations large distributed database • Scalability • Independent update • Robustness • Hierarchical database structure • Zones • How lookups are done • Caching/prefetching and TTLs • Reverse name lookup • What are the steps to creating your own domain? Lecture 13 15 -441 © 2008 77
- Slides: 69