Content Delivery and Remote DNS services

Slides:



Advertisements
Similar presentations
Dynamic Replica Placement for Scalable Content Delivery Yan Chen, Randy H. Katz, John D. Kubiatowicz {yanchen, randy, EECS Department.
Advertisements

EDNS0 Client-Subnet for DNS based CDNs
Page 1 / 14 The Mesh Comparison PLANET’s Layer 3 MAP products v.s. 3 rd ’s Layer 2 Mesh.
Computer Networks: Domain Name System. The domain name system (DNS) is an application-layer protocol for mapping domain names to IP addresses Vacation.
OSMOSIS Final Presentation. Introduction Osmosis System Scalable, distributed system. Many-to-many publisher-subscriber real time sensor data streams,
Flash Crowds And Denial of Service Attacks: Characterization and Implications for CDNs and Web Sites Aaron Beach Cs395 network security.
Anycast Jennifer Rexford Advanced Computer Networks Tuesdays/Thursdays 1:30pm-2:50pm.
1 Drafting Behind Akamai (Travelocity-Based Detouring) AoJan Su, David R. Choffnes, Aleksandar Kuzmanovic, and Fabian E. Bustamante Department of Electrical.
Content Delivery Networks. History Early 1990s sees 100% growth in internet traffic per year 1994 o Netscape forms and releases their first browser.
Hands-On Microsoft Windows Server 2008 Chapter 8 Managing Windows Server 2008 Network Services.
1 Content Distribution Networks. 2 Replication Issues Request distribution: how to transparently distribute requests for content among replication servers.
A User Experience-based Cloud Service Redeployment Mechanism KANG Yu.
Chapter 16 – DNS. DNS Domain Name Service This service allows client machines to resolve computer names (domain names) to IP addresses DNS works at the.
{ Content Distribution Networks ECE544 Dhananjay Makwana Principal Software Engineer, Semandex Networks 5/2/14ECE544.
Module 13: Network Load Balancing Fundamentals. Server Availability and Scalability Overview Windows Network Load Balancing Configuring Windows Network.
Ao-Jan Su, David R. Choffnes, Fabián E. Bustamante and Aleksandar Kuzmanovic Department of EECS Northwestern University Relative Network Positioning via.
CDN Brokering* Presented By Nick Arnold Authors Alexandros Biliris, et. Al.
1 On the Placement of Web Server Replicas Lili Qiu, Microsoft Research Venkata N. Padmanabhan, Microsoft Research Geoffrey M. Voelker, UCSD IEEE INFOCOM’2001,
Application of Content Computing in Honeyfarm Introduction Overview of CDN (content delivery network) Overview of honeypot and honeyfarm New redirection.
1 Chapter 7: NAT in Internet and Intranet Designs Designs That Include NAT Essential NAT Design Concepts Data Protection in NAT Designs NAT Design Optimization.
Web Caching and Replication Presented by Bhushan Sonawane.
Drafting Behind Akamai (Travelocity-Based Detouring) Dr. Yingwu Zhu.
On the Placement of Web Server Replicas Yu Cai. Paper On the Placement of Web Server Replicas Lili Qiu, Venkata N. Padmanabhan, Geoffrey M. Voelker Infocom.
John S. Otto Mario A. Sánchez John P. Rula Fabián E. Bustamante Northwestern, EECS.
Ip addressing: dhcp & dns
EDNS Client Subnet (ECS) in CDN solution
19 – Multimedia Networking
Scaling Network Load Balancing Clusters
Chapter 25 Domain Name System.
Chapter 9: Domain Name Servers
Module 8: Networking Services
Content Distribution Networks
CONNECTING TO THE INTERNET
Network Address Translation
Practical Censorship Evasion Leveraging Content Delivery Networks
Network Load Balancing
DNS.
VIRTUAL SERVERS Presented By: Ravi Joshi IV Year (IT)
Cisco OMD Feature Update
VIDIZMO Deployment Options
Introducing To Networking
Working at a Small-to-Medium Business or ISP – Chapter 7
LECTURE 34: WEB PROGRAMMING FOR SCALE
Net 323 D: Networks Protocols
Subject Name: Computer Communication Networks Subject Code: 10EC71
Working at a Small-to-Medium Business or ISP – Chapter 7
Chapter 5 Network and Transport Layers
ECE 671 – Lecture 16 Content Distribution Networks
HWP2 – Distributed search
AKAMAI INTELLIGENT PLATFORM™
LECTURE 32: WEB PROGRAMMING FOR SCALE
Working at a Small-to-Medium Business or ISP – Chapter 7
LECTURE 33: WEB PROGRAMMING FOR SCALE
Edge computing (1) Content Distribution Networks
Content Distribution Networks
AWS Cloud Computing Masaki.
HTTP and Abstraction on the Internet / The Need for DNS
Technical Capabilities
1 ADDRESS RESOLUTION PROTOCOL (ARP) & REVERSE ADDRESS RESOLUTION PROTOCOL ( RARP) K. PALANIVEL Systems Analyst, Computer Centre Pondicherry University,
SAMANVITHA RAMAYANAM 18TH FEBRUARY 2010 CPE 691
Domain Name System Refs: Chapter 9 RFC 1034 RFC 1035.
Ip addressing: dhcp & dns
AbbottLink™ - IP Address Overview
ONLINE SECURE DATA SERVICE
Computer Networks Primary, Secondary and Root Servers
EE 122: Lecture 22 (Overlay Networks)
An Engineering Approach to Computer Networking
Computer Networks Presentation
LECTURE 33: WEB PROGRAMMING FOR SCALE
Implementing Firewalls
Presentation transcript:

Content Delivery and Remote DNS services jinjian@cnnic.cn, May 11, 2013 Good afternoon, it is Much communication with chinacache Topic:Third-party dns affect cdn and further consideration

Remote DNS DNS :map names to addresses, or vice versa A large-scale distributed system CDNs: leverage DNS for dynamic routing Current CDN solutions assume proximity between users and their resolvers Use of remote DNS Servers concentrated farther from users Growing third-party DNS services Public DNS usage has grown to 11% of users! The web performance? Originally assume, their locations are very close in network, But recent years, Third-party are grow fast, for sp and end users, Apparently the third party dns can not guarantee So we start to think of under this condition

CDN deployment About 74% of the top 1000 web sites use CDNs This chart shows that the status ,blue, green, and The percentage of real traffic among these websites, pageviews, green bar, is it can be said it is very useful to improve the performance with a cdn solution By Alexa.com

CDNs depend on users DNS to redirect requests CDN: 27% annual increase So generally , First, for a cdn, content origin will synchronize other replica nodes, cdn will have a dynamic routing machanism, Then when an initiate a session using a local dns, directed to server , so if in theory With the growth of remote dns and cdn application, This kind of impact is becoming very phenomenal As a result, it will Cut down the effection of cdn Remote DNS services break this assumption

Current CDN redirections Depends on DNS location, CDN load balancing, network conditions Overlap zone between locations may be very small. According to such analysis, the cdn redirection will be decided by When an end user shift its dns from ,those possible destination of servers may have a little overlap if we do nothing on dns obviously

DNS extension Solution Main idea: directly provide client location to CDN Implemented as an extension: “edns-client-subnet” DNS resolver adds IP prefix of host to request CDN redirection based on location of host. EDNS effectiveness Approximate client location approach typically sufficient To solve this problem, dns extension solution is raised Main idea convey to cdn system, This way , dns resolver will It is eay to understand, If that, the effectiveness will be very sufficient So google company propose this

Performance improvement EDNS solution performance 45% performance improvement Curve graph shows the performance improvement, With the same condition, if we measure , there will a big improvement com Blue line vs red line Google DNS with ECS

E-DNS Both DNS and CDN services must support it only 1% of sites (top 1000) use a non-Google CDN that supports the extension. Google DNS was the only public service that supports the extension Why is it not ideal for CDNs to adopt E-DNS solution? This solution require that Currently only 1%. Some large CDNs make their redirection decisions by mapping DNS servers onto a set of\core points"in the network, which are in turn mapped to the CDN's infrastructure [4]; this initial mapping step is Very crucial. Under this approach, the problem is that the CDN cannot directly utilize the information provided by the extension. This issue presents a signicant barrier to adoption of the extension.

Clients based solution Main idea: move the resolver more closer to the user End host directly queries for CDN redirection. So CDN redirection can be based on client’s location Implementation: Run a DNS proxy on the user’s machine Monitor stream of requests to identify CDN redirections Use Direct Resolution to improve redirection quality Naturally, people move their attention to clients, This idea is propose in a public paper, The main idea is to make , that need to run Ref to: http://aqualab.cs.northwestern.edu/projects/namehelp

2 1 Direct Resolution Step 1: typical DNS query to recursive resolver Use recursive DNS to translate customer name to CDN Step 2: directly query CDN for an improved redirection 2 DR map 1 Initial map Step 1 canonical name record Step 2 with this canonical name then get an improved

Problems Driven by the active host proxy, this solution can not realized dynamic optimization, so it hasn’t played out the advantages of the CDN system. After the first packet, CDN system knows position information of all hosts , it is possible to actively change CDN replica for these hosts. How about “ passive host proxy , active CDN optimization”. The passive host proxy only receives commands of CDN system.

Passive host proxy , active CDN optimization Step 1: typical DNS query to recursive resolver After the first packet, CDN system knows the host’s IP. Initial map

Passive host proxy , active CDN optimization Step 2: CDN give an improved redirection to the passive proxy Initial map

Passive host proxy , active CDN optimization Step3: Update redirection Realize dynamic optimization. It is always Optimal Initial map

Other solutions? Better solutions: Don’t change DNS. Don’t change clients. It is an optimization problem of CDN. The CDN provider should take responsibility for this problem. A pure CDN based solution ? CCN? How?

The End. Thank you ! Thank you all for your time!