Aculab’s VoiceXML/CCXML interpreter software allows developers of communications solutions to implement standards-based interfaces to applications based on the Prosody family of media processing and signalling platforms. The use of interpreter software significantly speeds up the application development process by eliminating the need to develop with native product APIs, providing faster time to market, while reducing the associated costs.

Features and benefits Supports the W3C VoiceXML 2.0 and CCXML 1.0 standards for call scripting Easy integration with existing client-server and web services infrastructure Enables short development cycles – eliminates application development with telephony API Scalability for seamless growth from one to thousands of channels High performance guaranteed by pre-compilation of scripts and media caching Active call redundancy – automatic failover ensures no active calls lost Supports integration with 3rd party speech engines (TTS, ASR, SVI) ‘Future-ready’ architecture for video, IMS and mobility solutions Attractively priced – industry’s lowest cost per channel

Target applications
	IVR
	Voice portals
	Conferencing services
	Media servers
	Voicemail
	Unified messaging platforms
	Prepaid services

The VoiceXML interpreter conducts the call interaction with a caller, based on instructions from a VoiceXML script supplied by an application server. The interpreter processes documents (voice/call control applications) written in VoiceXML and CCXML languages and drives the underlying telephony platform, which is based on the Prosody family of products. The interpreter natively understands touchtone input and can manage pre-recorded audio prompts or files. It can also call upon voice technologies such as text-to-speech (TTS) and automatic speech recognition (ASR) for enhanced functionality.

Efficient applications
Aculab’s VoiceXML/CCXML interpreter enables the creation of a wide range of converged telephony solutions, operating in both VoIP and PSTN domains, which require interactive dialog between end users and high level applications. Current examples of voice-based applications include voicemail, interactive voice response, conferencing, prepaid services, voice portals and unified messaging platforms. Customers looking to develop multimodal communications solutions, with video, voice and data, will benefit from the interpreter’s ’future-ready‘ software architecture, which is already capable of handling multimedia call sessions.

Performance-optimised script execution algorithms guarantee uncompromised system operation for high density applications requiring thousands of simultaneous calls, for either hosted or on-premises deployment options. Aculab offers a highly efficient VoiceXML/CCXML interpreter, which guarantees operational performance. Most interpreters on the market today, as high level applications become more dense and complex and resource dependencies substantially increase the CPU and memory usage, experience performance bottlenecks. In contrast, Aculab’s interpreter remedies this situation by using a unique architecture that virtually eliminates application load and memory penalties, thereby allowing applications to stream smoothly.

Architecture
The architecture of Aculab’s VoiceXML/CCXML interpreter allows maximum flexibility for solution developers. Solutions can have either distributed architecture i.e., functional elements operate on separate hosts connected via IP, or a ’solution in a box‘, with all functional elements consolidated in one chassis.

The logical architecture of a VoiceXML/CCXML-based solution is shown in figure 1. The VoiceXML/CCXML interpreter communicates via web protocols (HTTP) to either a local or remote application server, which can query an enterprise’s middleware and database systems to dynamically drive content to the end user. The application server delivers pre-compiled VoiceXML/CCXML text pages and binary media files, which can be cached locally to improve runtime efficiency. The interactivity with the end user is allowed by providing the application server with spoken, touch-tone, and recorded input received back from the interpreter.

Although supplied in one software package, VoiceXML and CCXML parts of Aculab’s interpreter operate separately. The VoiceXML part is responsible for handling media interaction between high level applications and end users. It enables interactive dialogues with humans based on standard media processing features, including prompt recording and playback, music on hold and DTMF event handling. Integration with 3rd party speech engines is also supported, allowing solutions to utilise the benefits of interacting with end users using synthesised human speech and ASR. The CCXML part of the implementation is responsible for IP or TDM call control, allowing the initiation, management and termination of sessions with end user devices.

Figure 1: Aculab VoiceXML/CCXML solution architecture

VoiceXML applications can be housed either in hosted datacentres, or served directly from an enterprise application server in real-time. Prosody media processing platforms, which provide low level connectivity to IP and PSTN networks, utilise a highly bit-efficient, low latency ASSP protocol for communication with the high level control layer, consisting of VoiceXML/CCXML interpreter, TiNG, call control, and switch APIs. Intelligent caching of VoiceXML and audio files by the interpreter further minimises the impact on media latency. When a VoiceXML application server is located in a hosted datacentre, communication with backend enterprise databases typically occurs via real-time XML connections over a secure IP link. The integration with 3rd party voice engines (ASR, TTS, etc.) is enabled via Aculab’s MRCP client.

Active call redundancy
For solutions with high service continuity as the paramount objective, the VoiceXML/CCXML interpreter provides high resilience for system component failures. It supports sophisticated redundancy management to allow creation of applications with high availability, independent of the final solution scale or channel count. Utilising a layered software approach complemented by the latest CCXML specification, it possesses active redundancy, resulting in no live sessions being lost. This redundancy may be minimised or maximised, whereby a call can be sent back to step one of the script or sent back only one step, respectively. Many VoiceXML/CCXML systems on the market today rely on network level resilience. However, Aculab’s implementation is fully redundant, ensuring that live calls are never lost in the event of an application or system failure.

The basic requirement for the creation of resilient solutions is eliminating single points of failure by introducing redundant functional components. Aculab’s VoiceXML/CCXML interpreter can be distributed between several physical hosts and load balancing software may be implemented to minimise the impact of a failover event. Automatic failover mechanisms are also supported by the Prosody family of media processing platforms, making the solution failure tolerant from low level telephony hardware level up to the application part. An example of high availability solution is shown in figure 2.

Figure 2: Example architecture for high availability VoiceXML/CCXML solutions

Technical summary

Extensions to standard – Aculab has added custom objects within the VoiceXML syntax. Details can be found in the VoiceXML/CCXML Interpreter Software User’s Manual. Functionality includes:

	Manipulating a conference volume
	Performing SQL database operations
	Converting time values to a text string
	Performing file operations
	‘Sleeping’ for a fixed time period
	Mixing/filtering audio streams
	Playing tones
	Fax transmission (sending/receiving)
	Communication with the underlying system environment (telephony platform)
	Playing a file and allowing DTMF tones to replay and skip ahead/back (used for voicemail operations)

VoIP and PSTN call control – supports SIP and the full range of Aculab’s PSTN protocols and signalling systems (for more details please visit protocols and approvals.

Audio files – support for a wide set of audio file formats (please refer to the Prosody TiNG API), including the standard wave file (RIFF header format, filename suffix ’.wav’).

ASR/TTS voice engines – supported using Aculab’s MRCP software against any 3rd party voice engine supporting MRCP v2 (as specified in IETF draft: draft-ietf-speechsc-mrcpv2-12).

Scalability – Aculab’s VoiceXML/CCXML interpreter features linear scalability with no limit on the number of channels that are supported in a distributed environment.

Tested performance – developed for use in solutions that handle hundreds to thousands of simultaneous calls, standard testing of the interpreter set begins at 1000 channels per Dual Core Intel Xeon 5160 processor host running at 3GHz clock.

Operating system – Linux. For information on supported Linux distributions please refer to Aculab’s software support notes.

Telephony level integration – the interpreter supports all variants and form factors of Prosody X (PCI, cPCI and PCIe) and Prosody S at version 3.0 or later.

Administration and management tools – supported using SNMP protocol, which offers a standards-based capability for remote control and in-field service. Both Net-SNMP and Microsoft’s SNMP compatible MIBs are offered to provide choice.

Event logging and diagnostics – supports basic and verbose levels of event logging into file.

Licensing – the interpreter is available with a flexible software licence attached to a host machine. System channel counts may start form a single channel, with no fixed upper limit. Free development and evaluation licences are available (4 channels, 45 days).

Search our site

Email to a colleague

Print this page