Converged networks have come a long way since the 1990s. New applications like Instant Messaging, Unified Communications, and IP Telephony have accelerated the adoption and the deployment of converged voice and data services in the enterprise. Since its inception in the late 1990s, SIP (Session Initiation Protocol) has revolutionized the way people communicate with each other using converged services. SIP provides the framework for delivering voice, video, data, and wireless services seamlessly and transparently over a common network.
Enterprises are increasingly involved in mission-critical transactions and cannot afford infrastructure or services downtime. Enterprises need to provide the customer with the best possible user experience and need highly available and reliable services that meet desired SLAs (Service Level Agreements). This paper defines high availability and reliability in SIP-based services and discusses techniques of designing and building reliable SIP networks.
SIP high availability remains in its nascent stage, resulting in a lack of an industry standard for achieving SIP high availability and reliability. Until now, there has not been a reliable mechanism available to determine the state of a SIP Proxy or a Media server and when calls should be routed away from them. Calls get routed away after a server completely fails, and valuable time is spent probing and connecting to the backup server, which results in long service interruptions.
Most implementations for achieving SIP reliability are too complicated and proprietary, and do not guarantee the Quality of Service (QoS) assurance needed for voice/video traffic. They are also unable to guard and protect SIP components against security attacks that can bring a SIP resource or an entire site down.
SIP high availability is defined as the uninterrupted availability of core SIP components (SIP Proxy server, Registrar, and the Media server) that provide SIP services. SIP high availability and reliability is measured by the ability of the core SIP components to deliver high quality SIP services in the event of high call volume, link outages, device failures, and security attacks.
Consider the example in Figure 1, where employees at branch offices want to participate in a video conferencing session with employees located at the corporate office. The video conferencing session involves sending and receiving confidential and classified information in the form of voice, video, and text using SIP with Microsoft Instant Messenger or any other unified communication tool. Figure 1 illustrates the key places where the availability and reliability of SIP services could be compromised.
F5’s BIG-IP® platform provides all the necessary building blocks required to achieve total availability and reliability in a SIP-enabled network.Figure 2 illustrates how the example above can be transformed into a highly available and reliable SIP network using F5’s BIG-IP solution.
F5’s BIG-IP Link Controller™ (LC) and F5’s FirePass® secure remote access solution provide unparalleled protection from security attacks against SIP services and the site resources. The BIG-IP LC and LTM provide protection against common attacks against SIP services including DoS, DDos, IP Spoofing, and SYN flood. It also monitors the ports constantly, denies any illegal access attempts that prevent unauthorized monitoring of RTP traffic, and hides the ports on which the services are running. Using the SIP profile and iRules, white-lists could be used to enforce a strict policy of authorized SIP users. To maintain security, SIP traffic can be implemented using Transport Layer Security (TLS) or Secure Socket Layers (SSL).
The FirePass controller provides SSL VPN termination capabilities that enable SIP traffic to be encrypted and transported via an SSL VPN tunnel. In addition, FirePass could be configured to connect with a remote IPSec server to tunnel SIP. The FirePass controller also provides powerful client authorization and authentication capabilities like LDAP, Radius, etc., to allow secure access to users. Site-to-site remote connections can also be connected using F5 BIG-IP® WAN Optimization Module™ (WOM), which would provide a complete SSL solution.
F5’s BIG-IP Link Controller provides high availability and reliability to sites with multi-homed networks. It monitors the health and availability of links, detects errors across them, and transparently directs traffic across available links. By monitoring and managing bi-directional traffic to the site, the users always remain connected, ensuring high availability.
F5’s BIG-IP Global Traffic Manager (GTM) global load balancing solution provides high availability and reliability to a site. SIP Proxy servers use DNS name resolution to resolve SIP URLs. In the event of a site outage, the BIG-IP GTM directs users transparently to other available sites. Using its sophisticated health check system that includes TCP, SNMP, and HTTP checks, the BIG-IP GTM constantly monitors the state of a site and intelligently routes users to the best performing site when a site becomes unreliable or unavailable. BIG-IP GTM can also distribute SIP and RTP traffic across multiple sites using intelligent load balancing modes like Ratio, Global Availability, Least Connections, Round Trip Time, and so on, so that SIP resources at any site do not become overburdened or unreliable.
The BIG-IP Link Controller also ensures the service quality of SIP and RTP traffic by intelligently routing traffic over the best link based on QoS parameters like round trip time, completion rate, hops, and so on. It dynamically monitors the state of the links and routes traffic over the best link that guarantees the desired QoS for the SIP traffic.
Voice and video traffic are extremely sensitive to delays. Voice reliability is characterized by the packet loss, number of calls that can be processed reliably per second, and packet delays. Video reliability, on the other hand, is characterized by a variable packet rate, as video traffic is extremely susceptible to variable packet delay.
The BIG-IP Link Controller provides SIP voice traffic service quality and reliability by enabling users to configure QoS mechanisms based on parameters like completion rate, round trip time, and number of hops. This provides the users the ability to direct SIP voice traffic over a link that meets the stringent QoS requirements of maintaining the least number of delayed or dropped packets.
The BIG-IP Link Controller provides SIP video traffic service quality and reliability by enabling users to configure QoS mechanisms based on parameters like round trip time and completion rate. This provides the users the ability to select a link that would guarantee the least variable packet rate and provide the best video quality.
BIG-IP LTM can provide different levels of quality based on a number of SIP variables within the SIP session. Also, BIG-IP LTM helps reduce the overall chattiness of SIP-based communications.
With F5’s array of capabilities, it is possible to provide highly available, reliable, and secure SIP communications without sacrificing Quality of Service. F5 expands the level of control administrators can implement on SIP-based networks without sacrificing speed, security, and availability, while providing simple easy to understand management interfaces.