Next Lecture : Computer Networking. Overview. Peer-to-Peer Networks: BitTorrent. P2P, DNS, and Web Required readings

Transcription

1 Next Lecture : omputer Networking L-17 P2P and DNS P2P, DNS, and Web Required readings Do incentives build robustness in ittorrent? 2 Sections of oraldn Optional readings DNSaching, Semantic Free Referencing 2 Overview P2P DNS oraldn -to- Networks: ittorrent ittorrent history and motivation 2002:. ohen debuted ittorrent Key motivation: popular content Popularity exhibits temporal locality (Flash rowds) E.g., Slashdot/Digg effect, NN Web site on 9/11, release of a new movie or game Focused on efficient fetching, not searching Distribute same file to many peers Single publisher, many downloaders Preventing free-loading 3 4 1

2 ittorrent: Simultaneous Downloading Divide large file into many pieces Replicate different pieces on different peers peer with a complete piece can trade with other peers can (hopefully) assemble the entire file llows simultaneous downloading Retrieving different parts of the file from different peers at the same time nd uploading parts of the file to peers Important for very large files 5 ittorrent: Infrastructure node Keeps track of peers participating in the torrent s register with the tracker registers when it arrives periodically informs tracker it is still there selects peers for downloading Returns a random set of peers Including their IP addresses So the new peer knows who to contact for data an have trackerless system using DHT 6 ittorrent: hunks ittorrent: Overall rchitecture Large file divided into smaller pieces Fixed-sized chunks Typical chunk size of 256 Kbytes llows simultaneous transfers Downloading chunks from different neighbors Uploading chunks to other neighbors Learning what chunks your neighbors have Periodically asking them for a list File done when all chunks are downloaded 7 8 2

3 ittorrent: Overall rchitecture ittorrent: Overall rchitecture 9 10 ittorrent: Overall rchitecture ittorrent: Overall rchitecture

4 ittorrent: Overall rchitecture ittorrent: Overall rchitecture ittorrent: hunk Request Order Which chunks to request? ould download in order Like an HTTP client does Problem: many peers have the early chunks s have little to share with each other Limiting the scalability of the system Problem: eventually nobody has rare chunks E.g., the chunks need the end of the file Limiting the ability to complete a download Solutions: random selection and rarest first ittorrent: Rarest hunk First Which chunks to request first? The chunk with the fewest available copies I.e., the rarest chunk first enefits to the peer void starvation when some peers depart enefits to the system void starvation across all peers wanting a file alance load by equalizing # of copies of chunks

5 Free-Riding Problem in P2P Networks Vast majority of users are free-riders Most share no files and answer no queries Others limit it # of connections or upload speed few peers essentially act as servers few individuals contributing to the public good Making them hubs that basically act as a server ittorrent prevent ent free riding llow the fastest peers to download from you Occasionally let some free loaders download it-torrent: Preventing Free-Riding has limited upload bandwidth nd must share it among multiple peers Prioritizing the upload bandwidth: tit for tat Favor neighbors that are uploading at highest rate Rewarding the top few (e.g. four) peers Measure download bit rates from each neighbor Reciprocates by sending to the top few peers Recompute and reallocate every 10 seconds Optimistic unchoking Randomly try a new neighbor every 30 seconds To find a better partner and help new nodes startup ittyrant: Gaming ittorrent Lots of altruistic contributors High contributors take a long time to find good partners ctive sets are statically sized uploads to top N peers at rate 1/N E.g., if N=4 and peers upload at 15, 12, 10, 9, 8, 3 then peer uploading at rate 9 gets treated quite well ittyrant: Gaming ittorrent est to be the N th peer in the list, rather than 1 st Distribution of W suggests 14K/s is enough Dynamically probe for this value Use saved bandwidth to expand peer set hoose clients that maximize download/upload ratio Discussion Is progressive tax so bad? What if everyone does this?

6 Overview P2P DNS oraldn Obvious Solutions (1) Why not centralize DNS? Single point of failure Traffic volume Distant centralized database Single point of update Doesn t scale! Obvious Solutions (2) Why not use /etc/hosts? Original Name to ddress Mapping Flat namespace /etc/hosts SRI kept main copy Downloaded regularly ount of hosts was increasing: machine per domain machine per user Many more downloads Many more updates Domain Name System Goals asically building a wide area distributed database Scalability Decentralized maintenance Robustness Global scope Names mean the same thing everywhere Don t need tomicity Strong consistency

7 DNS Records DNS Design: Hierarchy Definitions RR format: (class, name, value, type, ttl) D contains tuples called resource records (RRs) lasses = Internet (IN), haosnet (H), etc. Each class defines value associated with type FOR IN class: Type= name is hostname value is IP address Type=NS name is domain (e.g. foo.com) value is name of authoritative name server for this domain Type=NME name is an alias name for some canonical (the real) name value is canonical name Type=MX value is hostname of mailserver associated with name org net edu root com uk gwu ucb cmu bu mit cs ece cmcl 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.cmcl.cs.cmu.edu Fred.cs.mit.edu DNS Design: Zone Definitions DNS Design: ont. org net edu root com uk gwu ucb cmu bu mit cs ece cmcl ca Zone = contiguous section of name space E.g., omplete tree, single node or subtree zone has an associated set of name servers Subtree Single node omplete Tree Zones are created by convincing owner node to create/delegate a subzone Records within zone stored multiple l redundant d 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 TP to ensure reliability Example: S.MU.EDU created by MU.EDU administrators

8 Servers/Resolvers Each host has a resolver Typically a library that applications can link to Local name servers hand-configured (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 DNS: Root Name Servers Responsible for root zone pprox. dozen root name servers worldwide urrently {a-m}.rootservers.net Local name servers contact root servers when they cannot resolve a name onfigured with wellknown root servers Lookup Methods Recursive query: Server goes out and searches for more info (recursive) Only returns final answer or not found Iterative query: Server responds with as much as it knows (iterative) I don t know this name, but ask this server Workload impact on choice? Local server typically does recursive Root/distant server does iterative 3 local name server dns.eurecom.fr 1 8 requesting host surf.eurecom.fr root name server iterated query intermediate name server dns.umass.edu 5 6 authoritative name server dns.cs.umass.edu gaia.cs.umass.edu Workload and aching What workload do you expect for different servers/names? Why might this be a problem? How can we solve this problem? DNS responses are cached Quick response for repeated translations Other queries may reuse some parts of lookup NS records for domains DNS negative queries are cached Don t have to repeat past mistakes E.g. misspellings, search strings in resolv.conf ached data periodically times out Lifetime (TTL) of data controlled by owner of data TTL passed with every record

9 Typical Resolution Subsequent Lookup Example root & edu DNS server ftp.cs.cmu.edu root & edu DNS server lient Local DNS server ns1.cmu.edu DNS server ns1.cs.cmu.edu DNS server lient Local DNS server cmu.edu DNS server cs.cmu.edu DNS server 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 TP? Try alternate servers on timeout Exponential backoff when retrying same server Same identifier for all queries Don t care which server responds Overview P2P DNS Server selection oraldn

10 DN Replicate content on many servers hallenges How to replicate content Where to replicate content How to find replicated content How to choose among known replicas How to direct clients towards replica DNS, HTTP 304 response, anycast, etc. kamai Server Selection Service is replicated in many places in network How to direct clients to a particular server? s part of routing anycast, cluster load balancing s part of application HTTP redirect s part of naming DNS Which server? Lowest load to balance load on servers est performance to improve client performance ased on Geography? RTT? Throughput? Load? ny alive node to provide fault tolerance Naming ased lient does name lookup for service Name server chooses appropriate server address What information can it base decision on? Server load/location must be collected Name service client Typically the local name server for client Round-robin Randomly choose replica void hot-spots [Semi-]static ]tti metrics ti Geography Route metrics How well would these work? How kamai Works lients fetch html document from primary server E.g. fetch index.html from cnn.com URLs for replicated content are replaced in html E.g. <img src= > replaced with <img src= > lient is forced to resolve axyz.g.akamaitech.net hostname

11 How kamai Works How is content replicated? kamai only replicates static content Serves about 7% of the Internet traffic! (in 2003) Modified name contains original file kamai server is asked for content First checks local cache If not in cache, requests file from primary server and caches file How kamai Works Root server gives NS record for akamai.net kamai.net name server returns NS record for g.akamaitech.net Name server chosen to be in region of client s name server TTL is large G.akamaitech.net nameserver choses server in region Should try to chose server that has file in cache - How to choose? Uses axyz name and consistent hash TTL is small How kamai Works kamai Subsequent Requests cnn.com (content provider) DNS root server kamai server cnn.com (content provider) DNS root server kamai server Get foo.jpg 11 Get 10 index. html End-user Get /cnn.com/foo.jpg kamai high-level DNS server kamai low-level DNS server kamai server Get index. html 1 2 kamai high-level DNS server End-user Get /cnn.com/foo.jpg kamai low-level DNS server kamai server

12 Overview P2P DNS oraldn Slides Patrick 65 12