T. Brisco, "DNS Support for Load Balancing", RFC 1794, Apr. 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....backup server tasks among all the hosts. Hence, each server host will serve as the primary for some requests and the backup for others [3, 10] We refer to this scheme as dual role servers. The distribution of requests can be done using standard load balancing techniques, such as Round Robin DNS [4] or a centralized load balancer (with failover capability) 5] 0 2 4 6 8 10 12 P U C y c l e s . Duplex Total Duplex pri Duplex bu Unreplicated Figure 2: Server hosts CPU cycles (in ....

T. Brisco, "DNS Support for Load Balancing," IETF RFC 1794 (April 1995).

....system. Section 4 describes the implementation of NetAirt as a module for Apache. Finally, Section 5 discusses our performance measurements, and Section 6 concludes. 2 Related Work The idea of exploiting DNS for client redirection emerged in 1995, yet originally only for load balancing [3]. In essence, a DNS server can be modified to respond to DNS queries concerning a given domain name with IP addresses of different machines. The only constraint is that any of these machines can service client requests targeting the service. For example, a query regarding www.globule.org is ....

T. Brisco, DNS support for load balancing, RFC1794, IETF, April 1995.

....Database Figure 1. Architecture of a scalable Web site. Requests are directed from the load balancer to one of several Web servers. The Web servers may access one or more databases for creating content. Web server. DNS servers can select one of these servers using a policy such as round robin [10]. There are other approaches which can be used for DNS load balances which offer some advantages over simple round robin [13] The DNS server can use information about the number of requests per unit time sent to a Web site as well as geographic information. The Internet2 Distributed Storage ....

T. Brisco. DNS Support for Load Balancing. IETF RFC 1794, April 1995.

....several approaches have been proposed. These can fall into the following two types. One is end to end approach and the other is network support approach. In the end to end approach, end system selects a good server to be accessed when a client sends a request. Content routing using DNS server[3][4] is the most well known approach of this category. A DNS server has information of candidate servers to be accessed and selects an adequate server when a client requests an IP address of a server. For server selection policy, roundrobin is used generally in a DNS approach. Network support ....

Thomas P. Brisco, "DNS support for load balancing," Internet Request for Comments (RFC 1794), April 1995.

.... request distribution 2.3.1 DNS Round Robin In a Server selection system that uses DNS, the DNS server will resolve the name of the service or domain to a different IP address using different algorithms such as load average of a simple round robin. This mechanism is specified in RFC 1974 [31]. DNS is the naming system used in the Internet, as such it supports massive scalability by making use of hierarchical organisation of naming information and caching at several levels of the naming system. DNS based server selection techniques base their effectiveness on setting a very low (down ....

T. Brisco. DNS Support for Load Balancing, RFC 1794. April 1995. http://www.rfc-editor.org/rfc/rfc1794.txt.

....the scheme, not to OS differences. The large overhead of the proxy approach is caused by the extraneous system calls and message copying that are necessary for moving the messages between the two levels of proxies and the server. V. RELATED WORK Early work in this field, such as Round Robin DNS [11] and DNS aliasing methods, focused on detecting a fault and routing future requests to available servers. Centralized schemes, such as the Magic Router [4] and Cisco Local Director [12] require request packets to travel through a central router where they are routed to the desired server. ....

T. Brisco, "DNS Support for Load Balancing," IETF RFC 1794 (April 1995).

....the beginning of the Web transaction when the name of the Web site must be mapped to one IP address of a component of the Web system. Through this simple mechanism, the authoritative DNS server (A DNS) for the Web site can select a different server for every address resolution request reaching it [19]. The A DNS replies to address requests with a tuple IP address, TTL , where the first entry is the IP address of one of the nodes in the distributed Web server system, and the second entry is the Time To Live (TTL) denoting the period of validity of the name address mapping that can be cached in ....

T. Brisco. DNS support for load balancing. RFC 1794, Apr. 1995.

....(MX) record for any host name ending in .cs.bell labs.com, even those that do not exist. Additionally, responses can be tailored according to factors other than the actual query. For example, it is standard practice to randomly order the results of a query to provide approximate load balancing [2]. As another example, content distribution networks like Akamai use custom DNS responses both for real time load balancing and to route clients to nearby servers [7] The system we have described can provide none of these capabilities, which depend on the coupling of the administrative hierarchy ....

T. Brisco. DNS support for load balancing. RFC 1794, April 1995.

....(MX) record for any host name ending in .cs.bell labs.com, even those that do not exist. Additionally, responses can be tailored according to factors other than the actual query. For example, it is standard practice to randomly order the results of a query to provide approximate load balancing [2]. As another example, content distribution networks like Akamai use custom DNS responses both for real time load balancing and to route clients to nearby servers [7] The system we have described can provide none of these capabilities, which depend on the coupling of the administrative hierarchy ....

T. Brisco. DNS support for load balancing. RFC 1794, April 1995.

....other policy. The performance of these SSI clustering mechanisms and policies is an area of research interest. A typical mechanism used for centralized dispatching is a Domain Name System (DNS) server that resolves a single host name into multiple IP addresses in round robin fashion. RFC 1794 [7] describes round robin DNS. Another mechanism is to have a single IP addresses spoof for multiple addresses. The Network Address Translation (NAT) standard [12] specifies this type of operation. The dispatcher is the single point of failure for a cluster. To improve reliability, standby ....

T. Brisco, "DNS Support for Load Balancing," RFC 1794, April 1995.

....been a lot of research done in the area of load balancing and request routing inside a single web server cluster. Several works analyze content based request distribution policies [PAB 98] ZBC 99] Others study mechanisms for redirecting requests to a particular web server, including DNS based [Bri95], TCP splicing [DKM96] CRS99] YL99] TCP hand off [ADZ99] PAB 98] and distributed redirection [ASD 00] There are multiple studies that analyze the performance of web servers and the effects of different design choices. Pai, et al. looked into performance of different server architectures in ....

T. Brisco. DNS support for Load Balancing. Technical Report RFC

....figures show that the processing overhead increases linearly in respect to message size in both cases and therefore both systems scale in the same manner with respect to message size and number of TCP IP packets. 5. Related Work If server replicas are running on multiple hosts, Round Robin DNS [8] and DNS aliasing methods can be used to provide fault tolerance for Web service by changing the host name to IP address mapping depending on the state of the system. When a server failure is detected, the advertised host name is no longer mapped to the IP address of the failed server host. As a ....

T. Brisco, "DNS Support for Load Balancing," IETF RFC 1794 (April 1995).

....between names and nexthop information) and forwards the message accordingly. This complements our work by providing the necessary functionality for implementing the service advertisement and discovery protocol. Work on resource discovery includes distributed name server such as DNS and X. 500 [8, 9, 17, 18]. More recently, there has been activity in the IETF to define protocol for locating services such as printers, disk server, etc. on local area networks. The Network Self managemenT ORganization (NESTOR) project at Columbia University [10] is an example of a directory based resource discovery ....

T. Brisco, "DNS support for load balancing," April 1995. RFC1794.

....Connections The throughput can be improved by replicating a proxy server on multiple processors, so that connections can be dynamically mapped to replicated proxy servers. Example: Round Robin DNS A well known example for dynamically distributed connections is a Round Robin extension to DNS [Brisco 95] All names of replicated WWW servers which shall share the connections are registered with a CNAME for the virtual WWW server. After each lookup the DNS server rotates the list of CNAMEs. The next client requesting the name of the WWW server will receive a different answer from the DNS ....

Thomas P. Brisco. "DNS Support for Load Balancing". RFC 1794, April 1995. (ftp://ds.internic.net/rfc/rfc1794.txt)

....by accessing files from either the local disk or remote disk through the network file system and transmitting them to the client. We assume that a client can send a request to any of the HTTP servers directly by using one of the routing mechanisms (such as the Domain Name Server s redirection [8], ONE IP mechanism [10] and router based redirection [11] s 1 s 2 s 3 s 4 s 5 QoS Web Server Messages Q Q Q Q Q Figure 1: Architecture of a QoS Web Server The primary goal of a QoS Web Server is to serve a file request only if servicing the requests does ....

Brisco, T. DNS Support for Load Balancing. Network Working Group, RFC 1794 http://andrew2.andrew.cmu.edu/rc/rfc1794.html.

.... the identity of servers that provide a specific service (e.g. LDAP) using a specified protocol (e.g. TCP) in a specified domain (e.g. service.com) 26] Earlier work suggests using the existing DNS zone transfer mechanism as a way to add flexible load balancing capability to a nameserver [27]. Finally, some recent work has proposed new mechanisms to reduce client latency related to name resolution using prefetching or proactive cache management techniques [28] 7] This work further affirms that DNS caching plays a crucial role in determining client perceived latency. VII. ....

Thomas P. Brisco, "DNS support for load balancing," Internet Request for Comments (RFC 1794), April 1995.

....(DNS) resolves to multiple IP addresses, DNS returns this list of IP addresses, rotated circularly after each request. If clients use the first address on this list, requests will be sent to the various IP addresses in round robin fashion 2 , thereby balancing the number of requests sent to each [15]. The NCSA scalable web server configuration [16] is one example of such use of DNS. When all servers are up, a true Round Robin scheme maps the n th request sent to the cluster to the n th server (modulo m) Thus, F(rn ) n (mod m) To get an ordered list for robustness, this is logically ....

T. Brisco. DNS support for load balancing, April 1995. RFC-1794.

.... [2, 4, 8, 15, 23, 24, 31, 32, 34, 35, 39] and it is now common for workstation clusters to include production load sharing programs such as LSF [40] or DQS [16] In addition, many network DNS servers, routers, and switches include the ability to multiplex incoming requests among equivalent servers [1, 7, 14], and several run time systems for distributed parallel computing on clusters or metacomputers include modules to balance requests among nodes [18, 21] Server load may also be combined with locality information for wide area network (WAN) information systems such as selecting an HTTP server or ....

T. Brisco. DNS Support for Load Balancing. Technical Report RFC 1794, Network Working Group, April 1995.

.... [2, 6, 10, 18, 19, 25, 26, 28, 29, 33] and it is now common for workstation clusters to include production load sharing programs such as LSF [34] or DQS [11] In addition, many network DNS servers, routers, and switches include the ability to multiplex incoming requests among equivalent servers [1, 5, 9], and several run time systems for distributed parallel computing on clusters or metacomputers include modules to balance requests among nodes [13, 16] Server load may also be combined with locality information for wide area network (WAN) information systems such as selecting an HTTP server or ....

T. Brisco. DNS Support for Load Balancing. Technical Report RFC 1794, Network Working Group, April 1995.

....locality and load balance. The Dispatch product by Resonate, Inc. supports a strategy that is also based on content based request distribution [21] Looselycoupled distributed servers are also widely deployed in the Internet and use various techniques for load balancing including DNS round robin [8], HTTP client re direction [1] Smart clients [27] source based forwarding [12] and hardware translation of network addresses [11] Load balancing strategies that seek to satisfy response time based performance criteria are discussed in [22] and the citations therein. Our work focuses on resource ....

T. Brisco. DNS Support for Load Balancing. RFC 1794, Apr. 1995.

....location mechanism. This has an inherent fault tolerance flaw in that the location of the service is hard coded into its name, thereby making it difficult to move the service to a new location when a fault occurs. A number of techniques have been developed to address this problem. DNS round robin [13] and specialized routers [16, 5] can be used to hide a number of replicated and or distributed servers behind a single logical name. Smart clients [50] can assist the service in transparent service location in the presence of faults by relying on browser support through Java and JavaScript. ....

....the node itself has crashed. This scenario violates the redundancy principle of clusters, and there is nothing the manager can do until the node comes back up. In many situations, we may be able to hide the affinity through other means. Affinity to IP addresses can be hidden via DNS round robin [13] or specialized routers [16, 5] Disk affinity may be reduced by using a shared disk across multiple nodes. 3.4.3 SNS Manager Failure In Section 2.2, we emphasized the use of a centralized SNS Manager to simplify the design of the system. In this section, we show how our strategies ensure that ....

Brisco, T. DNS Support for Load Balancing, Apr. 1995. RFC-1764.

....to systems which address similar problems in other contexts. We present our conclusions in section 7. 2 Alternative Solutions In this section, we will describe approaches for adding transparent fault tolerance and load balancing to Web services. Popular approaches currently include DNS Aliasing [Brisco 1995, Katz et al. 1994] HTTP redirect [Berners Lee 1995] Magic Routers [Anderson et al. 1996] fail safe TCP [Goldstein Dale 1995] and Active Networks [Wetherall Tennenhouse 1995] Figure 1 describes how Smart Clients fits in the space of existing solutions. We will describe each of the ....

T. Brisco. "DNS Support for Load Balancing", April 1995. Network Working Group RFC 1794.

....to ping either the router or the entire subnet. If it does not get a response, it assumes an interface failure and may shut down the host after sending an alert. Also, monitors can use redundant interfaces to avoid this scenario. Load Balancing We balance load in two ways: using round robin DNS [6] and DNS zone modification. If load increases beyond a high water mark, the host withdraws itself from the DNS zone, i.e. stops advertising itself as a service provider. This initiates a shift of incoming requests to other machines. Finally, when the load reduces below a low water mark, the host ....

....failures. Long et al. 17] have also discussed similar experiments to estimate Internet host availability but they do not classify the failures. Early solutions to the availability and loadbalancing for cluster based services consisted of roundrobin (RR) and load balancing (LB) extensions to DNS [6]. 7 Smith [24] and Kwan et al. 16] describe 7 Many popular web sites, e.g. www.ncsa.uiuc.edu, home. netscape.com use these schemes. their experiences with RRDNS for long term loadbalancing in UK national Web cache and in NCSA Web server, respectively. They mention that removing failed ....

T. Brisco. DNS Support for Load Balancing. Network Working Group, IETF, 1995. RFC 1794.

....to eliminate copying and streamline small request handling. The Lava project uses similar techniques in a microkernel environment [22] Other approaches for increasing Web server performance employ multiple machines. In this area, some work has focused on using multiple server nodes in parallel [6, 10, 13, 16, 19, 28], or sharing memory across machines [12, 15, 21] 8 Conclusion This paper presents a new portable highperformance Web server architecture, called asymmetric multi process event driven (AMPED) and describes an implementation of this architecture, the Flash Web server. Flash nearly matches the ....

T. Brisco. DNS Support for Load Balancing. RFC 1794, Apr. 1995.

.... predict the response time of the servers and thus make a good choice of which server to select [10, 11] In server oriented approaches a dedicated machine (usually a switch or router) assigns incoming requests to a cluster of servers according to some local selection scheme (usually round robin) [1, 9]. While the latter approach due to the single common entry point is neither scalable nor faulttolerant the former suffers from the need to regularly update aging statistics. We are aware of three methods [7, 8, 15] that like our method use client server cooperation to level loads ....

T. P. Brisco. DNS support for load balancing. Internet Request for Comments 1794, Apr. 1995.

....problems of scale. Distributing requests among replicas is an interesting issue, especially since it involves the two, often conflicting, goals of distributing load and increasing client server proximity. A number of sites in the Internet are mirrored and a number of schemes like DNS round robin [6], Distributed Director [9] and others [7, 15, 29] address the issue of request distribution. However, the mirroring is done statically and most of the request distribution schemes are designed for a static set of replicas. We believe that static replication is not a very feasible solution. ....

....4 Request Distribution This section describes the request distribution algorithm. The challenge in designing the request distribution algorithm lies in combining the goal of load distribution with that of choosing a replica that is close to the client. Choosing replicas in a round robin manner [6] would distribute the load but neglect proximity. On the other hand, always choosing the closest replica [9, 20] could result in poor load distribution. We have explored two approaches to request distribution, both of which try to increase clientserver proximity and distribute load at the same ....

T. Brisco. DNS Support for Load Balancing. Network Working Group RFC 1794, April 1995.

....in end hosts and the network interconnect. The second major limitation is inflexibility within the infrastructure itself, making it difficult to deploy new network services and applications. For example, a number of schemes for replication and load balancing [Katz et al. 1994, Berners Lee 1995, Brisco 1995] caching [Gw This work was supported in part by the Defense Advanced Research Projects Agency (F30602 95 C 0014, F30602 98 1 0205) the National Science Foundation (CDA 9401156) Sun Microsystems, California MICRO, Novell, Hewlett Packard, Intel, Microsoft, and Mitsubishi. Anderson was also ....

T. Brisco. "DNS Support for Load Balancing", April 1995. Network Working Group RFC 1794.

....by being replicated manually by the service provider. Load balancing across the wide area is achieved by instructing users to access a particular mirror site based on their location. To distribute load across servers, techniques such as HTTP redirect [Berners Lee 1995] or DNS Aliasing [Brisco 1995, Katz et al. 1994] can be used to send user requests to individual machines. With HTTP redirect, a front end machine redirects the client to resend the request to an available worker machine. This approach has the disadvantage of either adding a round trip message latency to each request or of ....

T. Brisco. "DNS Support for Load Balancing", April 1995. Network Working Group RFC 1794.

....geographically distributed but replicated manually by the service provider. Load balancing across the wide area is achieved by instructing users to access a particular mirror site based on their location. To distribute load in the local area techniques such as HTTP redirect [6] or DNS Aliasing [11, 38] can be used to send user requests to individual machines. With HTTP redirect, a front end machine redirects the client to resend the request to an available worker machine. This approach has the disadvantage of adding a round trip message latency to each request. Further, the front end machine ....

.... 7, 17, 31, 47, 10, 48] process management across the local area [53, 22, 42, 28, 57, 14] wide area file systems [35, 15, 39, 52, 16, 33, 1] wide area security systems [20, 46, 49, 27, 12, 55, 26] and mechanisms for hiding fault tolerance, load balancing, and dynamic relocation from end users [31, 38, 6, 11, 2]. In this section, we focus our discussion on efforts to exploit computing and storage resources of the Internet. Perhaps the closest to our work is the recent Active Networks proposal [51] Active Networks moves computing into the Internet by modifying Internet routers to be dynamically ....

T. Brisco. DNS Support for Load Balancing, April 1995. Network Working Group RFC 1794.

.... prototype scalable HTTP server is described in [8, 10] The method they use, illustrated in Figure 1, consists of having a set of HTTP servers on nodes in a cluster, that use the Andrew File System (AFS) 6] for sharing a set of HTML documents, and using the round robin Domain Name Server (RRDNS) [2, 13] for distributing accesses among the nodes in the cluster. Essentially, clients are presented a single name associated with the set of HTTP servers, and the RR DNS maps this single name to the different IP address of the HTTP servers, in a round robin manner; thus different clients will (ideally) ....

T. Brisco. DNS Support for Load Balancing. RFC 1794, Rutgers University, April 1995.

.... [2, 6, 9, 16, 17, 23, 24, 25, 26, 30] and it is now common for workstation clusters to include production load sharing programs such as LSF [31] or DQS [10] In addition, many network DNS servers, routers, and switches include the ability to multiplex incoming requests among equivalent servers [1, 5, 8], and several run time systems for distributed parallel computing on clusters or metacomputers include modules to balance requests among nodes [12, 14] Server load may also be combined with locality information for wide area network (WAN) information systems such as selecting an HTTP server or ....

T. Brisco. DNS Support for Load Balancing. Technical Report RFC 1794, Network Working Group, April 1995.

....future work. 2. Related Work A number of schemes have been proposed that address IP service replication and load balancing of distributed servers. DNS Based Approaches: The Internet naming server provides support for load balancing by having a one to many mapping from hostnames to IP addresses [7]. Early on, a round robin DNS was used with success for the NCSA Web server [14] Replicating services using naming mechanisms has a number of disadvantages. First, substantial caching happens in the name resolution hierarchy of DNS. This adds inertia before the load balancing effect of ....

T. Brisco. DNS Support for Load Balancing. RFC 1794

....that clients can use to contact it. The mechanism used to locate the service depends largely on the service itself. Web based services use Uniform Resource Locators (URLs) as the primary service location mechanism. Although this has fault tolerance implications, techniques such as DNS round robin (Brisco 1995) and specialized routers (Cisco Systems 1996, Anderson 1995) can be used to hide a number of replicated and or distributed servers behind a single logical name. Smart clients (Yoshikawa, Chun, Eastham, Anderson Culler 1997) can assist the service in transparent service location in the presence ....

....to those nodes. In the extreme case of affinity for a single node, the manager may simply be unable to restart the worker if the node itself has crashed. In many situations, we may be able to hide the affinity through other means. Affinity to IP addresses can be hidden via DNS round robin (Brisco 1995) or specialized routers (Cisco Systems 1996, Anderson 1995) Disk affinity may be reduced by using a shared disk across multiple nodes. 3.4.3 SNS Manager Failure In Section 2.2, we emphasized the use of a centralized SNS Manager to simplify the design of the system. We now show how our strategies ....

Brisco, T. (1995), DNS Support for Load Balancing. RFC-1764.

....in end hosts and the network interconnect. The second major limitation is inflexibility within the infrastructure itself, making it difficult to deploy new network services and applications. For example, a number of schemes for replication and load balancing [Katz et al. 1994, Berners Lee 1995, Brisco 1995] caching [Gw This work was supported in part by the Defense Advanced Research Projects Agency (F30602 95 C 0014, F30602 98 1 0205) the National Science Foundation (CDA 9401156, CDA 9624082) Sun Microsystems, California MICRO, Novell, Hewlett Packard, Intel, Microsoft, and Mitsubishi. ....

T. Brisco. "DNS Support for Load Balancing", April 1995. Network Working Group RFC 1794.

....the conclusions. 2 Related Work A number of schemes have been proposed that address IP service replication and load balancing of distributed servers. DNS Based Approaches: The Internet naming server provides support for load balancing by having a one to many mapping from hostnames to IP addresses [6]. Early on, a round robin DNS was used with success for the NCSA Web server [13] Replicating services using naming mechanisms has a number of disadvantages. First, caching at clients must be kept limited in order to take the most advantage of the round robin effect. This puts substantial load on ....

T. Brisco. DNS Support for Load Balancing. RFC 1794

....by being replicated manually by the service provider. Load balancing across the wide area is achieved by instructing users to access a particular mirror site based on their location. To distribute load across servers, techniques such as HTTP redirect [Berners Lee 1995] or DNS Aliasing [Brisco 1995, Katz et al. 1994] can be used to send user requests to individual machines. With HTTP redirect, a front end machine redirects the client to resend the request to an available worker machine. This approach has the disadvantage of either adding a round trip message latency to each request or of ....

T. Brisco. "DNS Support for Load Balancing", April 1995. Network Working Group RFC 1794.

....and performance of interfaces for telnet, FTP, and a scalable chat service. Section 5 evaluates our requirements above in the context of the Smart Client architecture. Section 6 describes related work, and Section 7 concludes. 2 Alternative Solutions Existing architectures include DNS Aliasing [Brisco 1995, Katz et al. 1994] HTTP redirect [Berners Lee 1995] Magic Routers [Anderson et al. 1996] failsafe TCP [Goldstein Dale 1995] and Active Networks [Wetherall Tennenhouse 1995] Figure 1 describes how Smart Clients fits in the space of existing solutions. We will describe each of the existing ....

T. Brisco. "DNS Support for Load Balancing ", April 1995. Network Working Group RFC 1794.

No context found.

T. Brisco, "DNS Support for Load Balancing", RFC 1794, Apr. 1995.

No context found.

T. P. Brisco. DNS support for load balancing. Internet Request for Comments 1794, Apr. 1995.

No context found.

Brisco, T., "DNS Support for Load Balancing", RFC 1794, April

No context found.

T. Brisco, "DNS Support for Load Balancing," RFC 1794, IETF (April 1995).

No context found.

T. Brisco. DNS Support for Load Balancing. RFC 1794, Apr 1995.

No context found.

Brisco, T., "DNS Support for Load Balancing", RFC 1794, April 1995.

No context found.

T. Brisco. DNS support for load balancing. Request for Comments 1794, Rutgers University, New Brunswick, New Jersey, Apr. 1995.

No context found.

T. Brisco, "DNS Support for Load Balancing", RFC 1794, Apr. 1995.

No context found.

Thomas P. Brisco, "DNS Support for Load Balancing," IETF, RFC 1794, April 1995.

No context found.

T. Brisco. DNS support for load balancing, 1995. RFC 1794.

No context found.

T. Brisco (1995). "DNS Support for Load Balancing." Network Working Group RFC 1794

No context found.

Brisco, T.P. DNS Support for Load Balancing. Rutgers University. 1994

No context found.

T. Brisco. DNS Support for Load Balancing. RFC 1794, April 1995.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC