| Cisco Systems, NetFlow, Technical white paper (August 1997). |
....Web server and Web cache clusters [1] are commonplace in many large Web provider sites and Internet Service Provider (ISP) GigaPOPs. In a typical installation, the server and the cache clusters are front ended by layer 4 switches which distribute connections across the nodes in the cluster [5]. The main objective of these layer 4 switches is to balance load among the servers in the cluster. Since, layer 4 switches are content blind, this approach mandates that the nodes in the Web server cluster are either completely replicated or share a common file system. Besides the storage ....
Cisco Local Director. Technical White Paper, 1998. Cisco Systems. 18
....We also investigate the performance impact of SSL on large e commerce sites which use clusters of servers for reasons of scalability. In a typical installation, the server cluster is front ended by a dispatcher (Figure 8) responsible for distributing connections across the nodes in the cluster [6]. Most often the dispatcher is agnostic of SSL session level information. As a result, SSL connections that can potentially reuse session states are routed to di erent server nodes. Unless the server nodes share their session caches, dispatching leads to poor SSL session reuse eciency and ....
....SSL session reuse in large e commerce installations which use clusters of servers to handle millions of transactions every day. In a typical e commerce installation the server cluster is front ended by a dispatcher (Figure 8) responsible for distributing connections across the nodes in the cluster [6]. The dispatcher is often unaware of SSL session level information. As a result, a connection that can potentially reuse the 12 SSL session state on a server node is often routed to a di erent node leading to poor SSL session reuse eciency. To understand the problem better, we need to know how ....
Cisco Local Director. Technical White Paper, 1998. Cisco Systems.
....caches. Web server and Web cache clusters [3] are commonplace in many large Web provider sites and Internet Service Provider (ISP) GigaPOPs. In a typical installation, the server and the cache clusters are front ended by layer 4 switches which distribute connections across the nodes in the cluster [7]. The main objective of these layer 4 switches is to balance load among the servers in the cluster. Since, layer 4 switches are content blind, this approach mandates that the nodes in the Web server cluster are either completely replicated or share a common file system. Besides the storage ....
Cisco Local Director. Technical White Paper, 1998. Cisco Systems.
....as 5 UUNET, the pioneer in IP multicast deployment, currently adopts a flat rate pricing approach with a Pm P u ratio of #400 [26] But its president is also predicting an Internet wide switch to usage sensitive pricing in the near future. 6 One example is Cisco s NetFlow software [6], which can be configured to support various charging schemes such as QoS based charging and distance sensitive charging. PRICING MULTICAST COMMUNICATION 291 edge metering devices to capture the necessary information for accounting and billing purposes. It is worth reiterating that the ....
Cisco Systems, NetFlow, Technical white paper (August 1997).
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