Executive Summary
Challenge
Cisco WAAS Overcomes the WAN
• Centralize costly distributed IT capital resources such as servers and storage into the data center
• Improve throughput and delivery of applications and application data to the enterprise edge
• Increase efficiency and control bandwidth utilization for existing WAN connections
• Maintain remote-office user application performance expectations
• Robust application-specific and protocol-specific acceleration-Cisco WAAS mitigates application layer performance challenges such as latency and bandwidth utilization through protocol acceleration, read-ahead, operation batching, multiplexing, and safe caching. The result is full correctness with protocol specification, full coherency of data, and a dramatically improved user experience when compared with native WAN access.
• Advanced protocol-agnostic network compression-Cisco WAAS is capable of compressing data in-flight using long-lived compression techniques including standards-based compression and cross-protocol data suppression. The result is a significant minimization of network bandwidth consumption and improved application throughput.
• Network-friendly throughput-improvement technologies-Cisco WAAS provides optimizations to TCP to improve WAN utilization efficiency and handling of WAN conditions, including packet loss, congestion, and recovery. The result is that communicating nodes are shielded from problematic WAN conditions, and Cisco WAAS manages these conditions with optimizations to improve throughput, performance, and response times.
Figure 1. Cisco WAAS Hardware Family

• Cisco WAAS Transport Flow Optimization (TFO)-Cisco WAAS provides optimizations that help improve TCP behavior in problematic WAN conditions to alleviate challenges associated with packet loss, congestion, recovery, and long fan networks (LFNs). With Cisco WAAS TFO, communicating nodes are shielded from WAN conditions, and WAE devices manage WAN conditions on behalf of the nodes to ensure that available capacity can be used to advantage, the effect of packet loss and congestion is mitigated, and throughput is increased. TFO maintains packet-network friendliness and safe coexistence with other network nodes communicating using standard TCP implementations.
• Cisco WAAS Data Redundancy Elimination (DRE)-DRE is a bidirectional database of blocks of data seen within TCP byte streams. DRE inspects incoming TCP traffic and identifies data patterns. As patterns are identified and added to the DRE database, they can then be used in the future as a compression history, and repeated patterns are replaced with very small signatures that instruct the distant device how to rebuild the original message. With DRE, bandwidth consumption is minimized, and latency associated with transferring data is also minimized because fewer packets need to be exchanged. DRE maintains full application and protocol coherency and correctness because the original message rebuilt by the distant WAE is always verified for accuracy at multiple levels and is application independent. Patterns that have been learned from one application flow can be used when another flow-even if using a different application-is seen. DRE can provide from 2:1 to 100:1 compression based on application, data, and workload.
• Persistent Lempel-Ziv (LZ) compression-Cisco WAAS implements LZ compression with a connection-oriented compression history to further minimize the amount of bandwidth consumed by a TCP connection. Persistent LZ compression, which can be used in conjunction with DRE or independently, provides from 2:1 to 5:1 compression based on the application used and data transmitted, in addition to any compression offered by DRE.
• Application Traffic Policy (ATP)-ATP is a robust acceleration and optimization management tool gives administrators the flexibility and control necessary to configure how Cisco WAAS handles specific application protocols. Cisco WAAS ships with default policies for more than 150 different traffic types and more than 25 application groups, and administrators can easily modify these existing policies or create new policies to match other application flows found in their environment.
• Industry-leading Wide Area File Services (WAFS) Functionality-Cisco WAAS builds upon the robust WAFS capabilities provided by the Cisco WAFS Family. Through protocol-specific optimizations, safe caching and data validation, read ahead, prediction, write-behind, multiplexing, and pipelining, Cisco WAAS provides extensive file services acceleration for Common Internet File System (CIFS) clients at the network edge, and can safely overcome protocol-specific performance limitations such as latency, data transfer, and bandwidth consumption. With Cisco WAAS acceleration, remote-office users receive LAN-like access to centralized file-server data, and with disconnected mode of operation, continuous ability to read files during periods of prolonged disconnection. Cisco WAAS also provides Windows-compatible print services that support "point-and-print" and centralized driver distribution for environments that prefer branch office print services. For those environments where centralized print servers are preferred, Cisco WAAS can also optimize that traffic as well. Furthermore, application acceleration technologies can also take advantage of the throughput improvements and compression provided by Cisco WAAS WAN optimization components when information must be transferred or messages must be exchanged across the WAN.
• Extensible application platform-Cisco WAAS is designed to meet current and future application delivery and infrastructure consolidation challenges. The modular software architecture provided by Cisco WAAS allows for additional robust application-specific adapters, WAN optimization components, or transparent integration of other optimization layers, thereby providing investment protection.
• Deployment flexibility and availability-Cisco WAAS is the only application-delivery platform today that offers deployment flexibility, availability, and service transparency. Cisco WAAS integrates transparently with clients, servers, and the network to preserve application configuration and network features. Network interception technologies such as the Web Cache Communication Protocol Version 2 (WCCPv2), policy-based routing (PBR), and server-load-balancing (SLB) platforms such as the Content Services Module (CSM) and ACE Series Application Control Engine for the Cisco Catalyst® 6500 series switchfacilitate fail-through operation with high availability, scalability, and load balancing. Additionally, Cisco WAAS can be deployed physically in-path using a network interface card (NIC) with fail-to-wire capabilities. Cisco WAAS transparency helps IT organizations maintain capital and operational investment in value-added network features such as optimized routing and path selection, quality of service (QoS), Network-Based Application Recognition (NBAR), NetFlow, firewall policies, and others.
• Network transparency-Cisco WAAS provides transparent optimizations and preserves original packet header information critical to network feature operation, including the source and destination IP and TCP information, allowing intermediary routers, switches, and firewalls to continue to perform functions against optimized packets such as classification, prioritization, access control, queuing, control, NetFlow, and routing decisions. Simply put, Cisco WAAS provides transparency necessary to allow value-added features provided by Cisco IOS to continue to operate. When coupled together, Cisco WAAS and IOS provide the most feature-rich framework for providing an application-optimized network infrastructure.
• Industry-leading scalability-Cisco WAAS, coupled with the Application Control Engine (ACE) module for the Catalyst 6500 series switch, provides WAN optimization and application acceleration scalability - up to 16Gbps of throughput and 4 million concurrent TCP connections. Cisco ACE provides not only a scalable off-path mechanism for distributing load amongst Cisco WAAS WAEs in the data center, but also additional application optimization, availability, security, and virtualization features.
Figure 2. Cisco WAAS and IOS Provide Industry-Leading Framework for Application Optimization

Deployment Flexibility
• WCCPv2, which was originally developed by Cisco Systems®, facilitates the transparent integration of application acceleration technology into the network with high availability and load sharing. The Cisco WAE devices in a given location advertise their availability to the router (or switch, multiple routers, or switches can be used for network path high availability) and specify that TCP traffic should be forwarded to the WAE. When the WAE devices join the service group with the router, the router monitors traffic for flows that should be forwarded to the WAE instead of the original destination. As the WAE begins receiving traffic, it selectively applies optimizations and protocol-level handling based on the configured application policy. With WCCPv2, up to 32 WAEs can join a service group with up to 32 routers, and each receives a portion of the workload that would otherwise traverse the WAN unoptimized. If a WAE fails, surviving members assume the workload of the failed WAE. If all WAE devices fail in a given location, traffic is forwarded across the WAN in an unoptimized fashion until a WAE is recovered.
• PBR is another deployment option available for the Cisco WAAS and the Cisco WAE. With PBR, the network administrator can configure a WAE or multiple WAEs as a next-hop router for all or specific TCP traffic. As the router receives TCP traffic, it forwards the traffic to the WAE as the next-hop router, where optimizations are applied based on the configured application policy. Like WCCPv2, PBR provides transparent integration into the packet network, and also offers high availability to the remote office or data center by using another WAE defined as a next hop if a WAE fails. If all WAEs fail, the policy-based route is considered unavailable, and traffic is forwarded across the WAN in an unoptimized fashion until a WAE is recovered.
• Physical inline is a deployment option that can be leveraged when out-of-path interception and redirection mechanisms are not possible. Cisco WAEs can be configured with an optional 4-port Gigabit Ethernet adapter that provides fail-to-wire capabilities. With this card, Cisco WAAS can be deployed on WAEs that are physically in-path between two network elements (for instance, the switch and the router, or the switch and the firewall). With fail-to-wire capabilities, should a hardware, software, or power issue be presented, mechanical relays in the card automatically fail to a pass-through condition. This allows Cisco WAAS to be deployed in such a way that a disruption would not create a network outage scenario should a problem be encountered.
Device Autodiscovery
Figure 3. Cisco WAAS Auto Discovery Process-Requestor

Figure 4. Cisco WAAS Auto Discovery Process-Receiver

Cisco WAAS TFO
• Large initial windows-Cisco WAAS increases the initial TCP window after autodiscovery to help TCP connections more quickly exit slow-start so that WAN bandwidth can be used more quickly, providing performance improvement for not only short-lived connections that are normally starved for bandwidth, but also longer-lived connections that are forced to reenter slow-start because of congestion.
• Window scaling-Cisco WAAS allows devices using standard TCP implementations to enjoy the benefits of window scaling without client or server modification. Employing window scaling allows Cisco WAAS to drastically improve performance over LFNs, or networks that have high bandwidth and high delay characteristics. By safely scaling TCP windows, Cisco WAAS helps applications that would normally be throughput-constrained to perform well in WAN environments and take full advantage of the available capacity provided by the WAN.
• Advanced congestion management and loss recovery-Cisco WAAS uses advanced congestion-management and loss-recovery techniques to help ensure that maximum throughput is safely restored after scenarios in which packet loss is encountered. Cisco WAAS advanced congestion management not only helps improve overall throughput, but also maintains compatibility with other TCP implementations (including standard TCP implementations) that may be in use on the network.
Figure 5. Cisco WAAS TFO Improves Application Performance and Reliability

Advanced Network Compression
Figure 6. Cisco WAAS Advanced Compression

Application Traffic Policy Engine
Table 1. Default Cisco WAAS Application Policy: Common Application Types Optimized by Cisco WAAS
Industry-Leading Wide Area File Services
• Protocol-specific acceleration-Cisco WAAS supports all CIFS clients and dialects, and examines client-server communications on a message-by-message basis to fully understand the operations being performed. As such, Cisco WAAS can make intelligent decisions on how to most appropriately accelerate specific operations. Such acceleration can include read ahead, message and operation batching, multiplexing, or pipelining. In many cases Cisco WAAS can safely handle message traffic locally without compromising protocol correctness or data integrity, thereby providing LAN-like performance. In cases in which messages must traverse the WAN without modification to maintain correctness, coherency, security, or data integrity, Cisco WAAS transfers them using the underlying network optimizations, including TFO, DRE, and persistent LZ compression. With protocol-specific acceleration, Cisco WAAS provides LAN-like performance for remote users accessing centralized file server storage without compromising protocol semantics, coherency, or correctness. Furthermore, Cisco WAAS safe acceleration for file services protocols work in environments with global collaboration, including CAD/CAM, software development, and databases through global file locking, because file lock requests always propagate to the origin server.
• Investment protection-Cisco WAAS is built with a future-proof architecture such that should protocols change, Cisco WAAS can still safely apply acceleration. Cisco is a member of relevant partner programs with companies that develop the CIFS protocol to ensure timely improvement to Cisco WAAS to ensure support for future versions of the CIFS protocol.
• Safe Data and metadata caching-Along with protocol-specific acceleration, Cisco WAAS can also cache data and metadata when safe. By employing an application-specific data and metadata cache, Cisco WAAS can serve usable content, once validated for coherency, to the requesting user. For scenarios in which an object is cached but has been modified, Cisco WAAS can fetch the updated contents using network optimizations such as TFO, DRE, and persistent LZ compression. Under no circumstances can Cisco WAAS serve outdated or changed content, as the origin server owns the authoritative copy of files and also the state of each file, including file locks.
• Centralized file storage-Cisco WAAS allows IT to centralize distributed file servers, storage capacity, and data into the data center where IT staff is readily available. Centralizing distributed servers and storage has many tangible benefits, including:
– Fewer devices to manage-Cisco WAAS can effectively replace the need for distributed file servers, minimizing the number of devices to manage in the infrastructure and eliminating many costly components, including servers, server operating systems and maintenance, OS patching and hotfixes, antivirus, tape drives and libraries, tape cartridges, backup software, and more.
– Taking advantage of existing data center infrastructure-With a consolidated infrastructure, application and file servers can use the data center infrastructure components to full potential, including server virtualization and storage virtualization. With protocol-specific acceleration, most of the workload is handled at the edge by Cisco WAAS, thereby enabling greater economies of scale with existing server and storage infrastructure.
– Fewer points of data protection-By consolidating distributed file server storage and data into the data center, fewer copies of data must be protected, thereby helping control the cost of protecting data and maintaining compliance with federal or industry regulation.
– Streamlined disaster recovery and business continuity-Cisco WAAS facilitates consolidation and minimizes the amount of application and file storage infrastructure necessary to support a distributed enterprise. Fewer remote application instances and fewer copies of data significantly simplifies disaster recovery and business continuity planning, deployment, and management.
• Data integrity, correctness, and coherency-Cisco WAAS accelerates protocols only when it is safe. Critical messages, including authentication, authorization, file lock requests, and write requests, always propagate to the data center without modification by Cisco WAAS and can take advantage of the underlying WAN optimization framework provided through DRE, TFO, and persistent LZ compression. With Cisco WAAS, the data center file server or NAS device always owns the data itself, the state of the data, and any locks applied against the data. Therefore, when a user closes a file and exits the application, the data is safely stored in the data center.
• Integration with advanced network compression-Cisco WAAS takes advantage of the advanced compression layers provided by DRE and persistent LZ compression. By integrating with DRE and persistent LZ compression, Cisco WAAS dramatically minimizes the amount of bandwidth consumed by application messaging and data transfer to the amount of changed data. This feature is extremely helpful in cases in which messages must traverse the WAN or when file data is being written back to the file server unmodified or partially modified.
• Disconnected mode of operation-Cisco WAAS application acceleration also provides a read-only disconnected mode of operation for situations in which the WAN or the file server has gone offline for an extended period of time. For information that needs to be accessible during periods of disconnection, Cisco WAAS aggressively caches files, folders, metadata information, and access control information. During periods of disconnection, a nearby domain controller can be used to authenticate users and the WAE can validate that requesting users are authorized to access cached data in a read-only fashion.
Summary
