Home >> March 2010 Edition >> FOCUS — Military Satellite WANS—How To Optimize, Accelerate + Provide Interoperability
FOCUS — Military Satellite WANS—How To Optimize, Accelerate + Provide Interoperability
by Gordon Dorworth

stampede intro msm mar10
The Spirit Of Iridium poster by Robert T. McCall

There are many operational and environmental conditions which adversely affect military satellite communications, impairing network performance and consuming valuable bandwidth. While satellite communications enable mobility and communications in remote locations, satellite-based Internet connectivity suffers due to adverse weather, interference, high-latency and transport and application protocol inefficiencies.

stampede f1 msm mar10

Many of these factors degrade throughput, while others limit the amount of traffic that can run over a satellite link. Additionally, interoperability among disparate military networks can jeopardize mission-critical communications.

To combat this, NASA and the Department of Defense (DoD) jointly created the Space Communications Protocol Standards (SCPS) protocol suite that has been deployed throughout the U.S. military. They rely on SCPS-TP to overcome transport protocol interoperability, and enable satellite WAN links to have a greater amount of traffic flow. This helps deliver more reliable satellite links in support of military missions, regardless of the operating conditions.

In addition, the U.S. DoD operates in areas of the world where bandwidth is extremely scarce and expensive. The military is very mobile, and communications can be extremely time sensitive. Providing a comprehensive solution for reliable and fast military communications has become an ever-increasing challenge — balancing security, speed of delivery and bandwidth consumption.

There is an emerging need to combine SCPS-TP and robust WAN optimization and application acceleration features into a single, small footprint solution for military satellite communications. WAN optimization and application acceleration capabilities not only substantially reduce the amount of unnecessary traffic going over a satellite WAN, they also increase the throughput beyond traditional SCPS-TP solutions.

This article details the obstacles faced by military satellite communications, and how solutions that integrate SCPS-TP and WAN optimization are able to provide military and government organizations with greater use of bandwidth, and speedy, secure access to mission-critical information needed to do their job, no matter where they are located.

Two essential requirements for improving reliability, performance and interoperability.

The first requirement is transport protocol interoperability via SCPS-TP. As mentioned above, the SCPS-TP protocol suite was jointly developed by NASA and the DoD. However, an SCPS-TP implementation that has become the “defacto standard” transport protocol is SkipWare from Global Protocols, a protocol engineering firm specializing in the development and integration of bandwidth efficient, error-tolerant protocol solutions based on SCPS-TP. SkipWare is the industry’s first commercial SCPS-TP implementation.

The second requirement is acceleration and optimization technologies from companies such as Stampede Technologies that provide considerable performance and bandwidth availability improvements for Internet-based satellite communications. Satellite WANs require acceleration and optimization technologies to improve bandwidth utilization, in order to squeeze greater amounts of traffic into a satellite link, and accelerate traffic going over the link to deliver the applications more quickly to the users.

The diagram above shows satellite communications using SPCS-enabled WAN optimization controllers at the head-end and an appliance at one remote site. Software is deployed at a second remote site.

Many of the WAN optimization solutions on the market are not SCPS-TP-compliant, and therefore, they lack the degree of interoperability required by military personnel, while standalone SCPS-TP implementations have limited acceleration capabilities.

However, there are solutions on the market specifically designed to accelerate, optimize and provide interoperability for military satellite communications. These solutions offer the best of both worlds — rich optimization and acceleration capabilities, and integrated SPCS-TP.

WAN optimization solutions with integrated SPCS-TP provide comprehensive optimization and acceleration technologies that work together to address network bandwidth constraints, high-latency, traffic volumes, transport and application protocol inefficiencies, excessive content/object transfers, and application contentions. These solutions can save as much as 75 percent or more of bandwidth capacity, while dramatically increasing the response times experienced by end-users.

The benefits that can be achieved:
  • Obtain greater value from your bandwidth
  • Improve application delivery times
  • Improve WAN response times
  • Optimize connectivity with any SCPS-TP-based military network
  • Reduce the amount of data that is sent over the satellite link — allowing you to send more traffic
  • Reduce the number of round trips (for transport & applications) to complete a transaction
  • Offload tasks from clients and servers — allowing them to handle a greater amount of traffic
  • Secure data and transactions — end-to-end

Primary Issues

High Latency — Latency reduces the amount of data that can be transmitted through a network link, regardless of how much bandwidth is available. On terrestrial-based WANs, latency can range from 0.1ms to 200ms. However, on a satellite link, latency can be as high as 1,000ms. Because of latency, even though you pay every month for a certain amount of bandwidth, in reality, you may not be getting the full value of the connection. Latency has significant adverse affects on TCP and web-based applications that require extensive handshaking. For satellite communications, one major challenge with respect to the performance of Internet applications is the latency between two earth stations connected by a satellite. For GEO satellite communications systems, latency is at least 250ms. Beyond the issue of distance; framing, queuing, and on-board switching can make end-to-end latency as high as 400ms.

Transport Protocol Limitations + Inefficiencies — TCP has a maximum window size of 64 kilobytes. This requires that for each round trip, TCP only allows 64KB to be sent. For example, sending a 64MB file over a satellite WAN will require 1000 round trips in order to complete the transmission. TCP packets require an acknowledgement (Ack) that they were successfully received before additional packets can be sent. If the round trip time takes 100ms, then that transaction is going to take 100 seconds to complete — regardless of how much bandwidth the link is capable of supporting.

Chatty Application Protocols — Applications have their own protocols with similar challenges to TCP, and can be even more inefficient. For example, Web pages are comprised of many separate objects, each of which is requested by the user (browser) and retrieved sequentially. Similar to TCP, a browser will wait for a requested object to arrive before requesting the next one. Web applications can generate hundreds of round trips by themselves. These round trips are in addition to the round trips generated by TCP. Web applications can also have window sizes 15 to 20 percent smaller than TCP, requiring even more round trips.

Bandwidth Optimization — Satellite WAN links can be costly, especially in remote locations and foreign countries where network connectivity options are limited. Bandwidth optimization helps to squeeze more bandwidth into network links. However, as important as bandwidth optimization is, it is also vital to accelerate traffic throughput. Optimizing bandwidth doesn’t necessarily help with throughput, particularly on high-latency links and chatty applications.

TCP Acceleration — To address the inefficiencies of TCP, WAN optimization solutions manage network connections in several ways to optimize the flow of data and reduce the impact on the network, application servers and end user devices. WAN optimization solutions can maintain a consistent pool of connections (multiplexing) between themselves and the Web application servers. The Web application servers are offloaded from managing the connections, and are isolated from inadvertent session disconnects. Using WAN optimization client technology, a persistent connection between the client and the WAN optimization appliance is always maintained, even when browsers close and reopen sessions. These sessions are also multiplexed across multiple connections, improving throughput and response time. This persistent connection capability is extremely important for AJAX and Web 2.0 applications that constantly open and close sessions as they poll and access various Web services.

HTTP Acceleration — HTTP acceleration enables HTTP browser traffic to be intermixed across multiple “pipelines”. All browser activity is optimized, including the network-intensive polling associated with Web 2.0 and AJAX applications. A key advantage of using a WAN optimization client-side implementation is that communication resources can be shared across multiple applications, and all HTTP requests and responses from any application (including multiple browsers) that are intermixed simultaneously across multiple concurrent sessions. HTTP acceleration serves as a platform for the consolidation and aggregation of all web-based traffic from a given client. Much as physical “link aggregation” uses multiple Ethernet network cables/ports in parallel to increase network link speeds beyond the limits of any one single connection, HTTP acceleration logically aggregates multiple HTTP protocol streams across a few TCP sessions. Individual objects or pieces of objects can be split into any size and then multiplexed with other object data and reconstructed as needed, which makes HTTP acceleration ideal for converged networks and managed service networks that must deliver mixed payloads consisting of mission-critical applications and other network-intensive traffic.

Caching — Caching techniques are used to position frequently requested content and objects closer to the users requesting the data. Caching maintains copies of routinely accessed data to eliminate unnecessary requests to web servers. Cache differencing goes a step further and maintains identical copies of the browser’s cache on the local WAN optimization device. Intelligent cache differencing understands what data has actually changed, and then transfers only the changed data. The result is less data transferred, improved network utilization, and increased user productivity. An important aspect of cache differencing to be aware of is the ability to perform differencing not only on HTML GET requests, but also on POST requests. This is significant because responses to posts are always marked non-cacheable, and most applications that are based on SOAP and XML (including most AJAX applications) issue SOAP requests via the HTML POST command.

Pre-Caching helps eliminate network bottlenecks, and improve end-user response times. Administrators can define content that is automatically distributed (Pre-Cached) to specific users at off peak hours. During a predefined time, the end-user devices will randomly pre-cache the content, enabling instant access to the information when the user needs it.

Compression — Compression eliminates non-essential information from data that is moved from one location to another. The data is then reassembled to its original form after the transfer is complete. Compressing data reduces network traffic and accelerates the delivery of time-sensitive information. GZIP Compression is a standard technique used to compress data that is sent to browsers. This is useful for reducing the text portions of pages. However, GZIP is not typically used for attachment compression, or for inbound compression from the browser. GZIP cannot be used to compress HTTP headers, cookies or image data. To accomplish these tasks, bi-directional compression is used to reduce the data size through extremely efficient and intelligent compression techniques. When a WAN optimization solution is deployed in a two-sided environment (both at the head-end or NOC and the remote sites), bi-directional compression provides compression for HTTP headers, application cookies, text and data objects, JPEG files with image reduction, file attachments and file uploads and downloads.

Quality of Service (QoS) — QoS enables the network to use queuing functionality to provide preferential treatment to certain classes of traffic. Traffic Shaping ensures on-time delivery of time-critical information. Traffic shaping allows different TCP ports to be assigned to individual applications. Specific port assignments, priorities and policies can be assigned at the database-level, guaranteeing QoS for mission-critical applications.

stampede author msm mar10 Military satellite communications are not only vital for mission-critical operations, they are also important for deployed military and government personnel to be able to communicate with family and colleagues back home. Today’s satellites use Internet technology that is inherently hampered by TCP and application protocol inefficiencies, latency and adverse environmental conditions that can drastically impair network performance, reliability and bandwidth use. Additionally, open-source standards are critical for ensuring complete transport protocol interoperability among disparate military networks.

WAN optimization and application acceleration solutions that integrate SCPS-TP can substantially reduce the amount of unwanted and unnecessary traffic going over a satellite WAN, while increasing throughput. These solutions are able to provide military and government personnel with fast, reliable and secure access to mission-critical information — from anywhere in the world.