Australian businesses are already in the process of migrating to IP, with the Australian data services market expecting in the short-term to reach the important milestone of service revenue from IP-based services overtaking the service revenue of legacy technologies.
- Data services market overview
- The growth of Ethernet services
- Ethernet benefits
- Ethernet adoption drivers & inhibitors
- Current Ethernet offerings
- Enterprise cloud computing & Ethernet services
- Related thought leadership
Data services market overview
In 2008, the Australian retail data communications market was worth $2.99 billion, an increase of 1.8% from 2007.
Australian businesses are continuing to migrate from legacy data services to IP, with revenue generated by legacy technologies including leased private line, frame relay, ATM, and ISDN, dropping by 9.5% year-on-year to below $1.67 billion in 2008. As a result, legacy services share of total data services revenue reduced from 63% in 2007 to 56% in 2008. Within the legacy technology segment, ISDN data was the fastest declining category (-12.6%), with businesses increasingly moving away from dedicated circuits. ISDN is significantly more expensive to provision and deploy and generally offers lower bandwidth than alternative DSL/SHDSL/gSHDSL technologies. Its decline has been compounded by it lack of cost competitiveness.
In contrast, IP-based technologies have grown by 20.8% from $1.1 billion in 2007 to $1.3 billion in 2008. This growth was attributed to Ethernet and IP-VPN/MPLS aggressively replacing legacy networks since 2007.
Telsyte forecast that, in aggregate, the data services market will grow by an average growth rate of just over 3% per annum between 2009 and 2013. In 2009, revenue is expected to increase by 2% to $3.04 billion.
Between 2009 and 2011, Australian businesses will continue to migrate their data networks from legacy technology-based networks to next-generation, IP-based networks. The market share of the legacy segments is expected to shrink from 63% in 2007 to just over 20% of the total market by the end of 2013. Telsyte believes that ATM will be nearly completely phased out by the end of the forecast period. Where it does persist, it will be in transport links as it will have disappeared entirely as an access technology.
The completion of NBN will provide a platform that will deliver enhanced access bandwidths based on IP protocols, and consequently a more cost effective delivery of IP-based applications, which in turn is expected to drive demand.
With the decline in legacy technologies, future growth in the data services market will be solely driven by the IP-based segments of IP-VPN/MPLS, Ethernet and SIP trunking, and it is expected that IP based services will make up of 80% of data services market by the end of 2013 from its base of 44% in 2008.
Figure 1 – Australian data services market revenue forecast by generation, 2007-2013
The growth of Ethernet services
Telsyte observed that the Ethernet services market grew by 15% from 2007 to 2008, to reach $425 million. Telsyte predicts that the segment will experience double digit growth until 2011 and expects Ethernet service revenue to reach $680 million by the end of 2013. Service pricing is expected to come under pressure as the level of competition and bandwidth availability increases.
Figure 2 – Australian Ethernet services Market Revenue Forecast, 2007-2013
Just why is Ethernet services proving so popular and growing so rapidly? Two short words explain the phenomena: – standards and cost. And both are interrelated.
For the past 15 years, the rate of investment in application of IP protocols, and Ethernet in particular, to telecoms services has risen dramatically. This is because vendors can realise greater “bang for buck” out of product R&D focused on IP/Ethernet than on any of the legacy protocols. They are leveraging an already high volume low-cost manufacturing base that supports enterprise and consumer IT products in which Ethernet is already ubiquitous, and leveraging the broad publicly funded University research base, which by mid 90’s had focused almost exclusively on IP and Ethernet.
A key driver was that legacy services did not scale well in response to increased network size and increased bandwidths. The legacy access technologies like E1/T1, and ISDN were made to “fit” into the higher bandwidth transport protocols of SDH/SONET, with inefficiencies and with provisioning and operational complexities. As telecoms were still dominated by voice revenues at that time, the concepts of circuit switching (as compared to packet routing) were deeply embedded in the telco community thinking and in their revenue models.
Universities and vendors already recognised by then, even if the telco carriers had not, that ISDN was already too limited as an access technology, and that ATM based equipment and deployments were too complex, too costly, and too slow, with limited horizons for improvement in speed. IP Packet based technologies were cheaper and faster to develop and deploy, and in a rerun of the 1970’s style “Beta vs VHS wars”, IP was fast winning the hearts and minds of the technology community.
But IP Packet based equipment had one major hurdle to overcome, and while not yet solved, has been effectively addressed since the mid 2000’s in some application areas. That hurdle, Quality of Service (QoS), is the ability to provide different priority to different applications, users, or data flows, or to guarantee a certain level of performance to a data flow. It is fundamental to traffic that intermixes voice, video and data, where video in particular is very sensitive to jitter delay and lost packets, while voice quality can be obviously degraded by as little as a loss or delay of 20 msec worth of data.
QoS was an area where ATM and legacy protocols could claim clear superiority, as these could be enforced in hardware, as IP protocols had no clearly ascendant mechanism for delivering reliable QoS even within one carrier’s network, let alone across an Internet comprising of dozens or more of unrelated and unpredictable networks. Given that the majority of traffic until early this decade was voice, legacy protocols were considered “the gold standard” in delivering reliable service. To provide high quality communication over a best-effort network lacking a QoS mechanism, IP networks initially could only deliver by over-provisioning the capacity so that it is sufficient for the expected peak traffic load. This was a cost overhead that supported the arguments for legacy protocols for a long time.
Now with legacy protocols like ISDN and ATM locked in a time warp of bandwidth limitations, IP protocols have gained ascendancy by their ability to offer ubiquitous, seamless, standards based access and transport over speeds from dial-up to 100Gbps.
In the access arena though, the rapid rise of IP/Ethernet access has been deliberately slowed by telcos, anxious at the erosion of the profitable legacy protocol revenue base and uncertain whether the lower price-per-bit of data traffic would offset these losses, even with the high rates of growth in data volumes.
However, there was one more hurdle IP supporters faced to gain universal acceptance of IP’s dominance as both an access and as a transport protocol set. That was the ability of multiple technologies to provide IP access and to carry IP transport, which often involved expensive, and not always compatible, network elements to “groom” or interface the data flow.
There were multiple competing techniques for carrying Ethernet packets over telecoms services. These included:
- Pure Ethernet, i.e. Layer 2 VLAN switching only
- Ethernet over TDM (pseudo-wire, an access methodology)
- Ethernet over SDH
- Ethernet over MPLS
- Ethernet over Fibre, i.e Ethernet over CWDM/DWDM (Coarse Wave Division Multiplexing and Dense Wave Division Multiplexing)
Pure Ethernet-based deployments are cheap but less reliable, limited QoS and scalable, and thus are usually limited to small scale deployments.
SDH-based deployments are useful when there is an existing SDH infrastructure already in place, its main shortcoming being the loss of flexibility in bandwidth management due to the rigid protocol hierarchy.
MPLS based deployments are costly but highly reliable and scalable, and are typically used by large service providers.
In 2001 the Metro Ethernet Forum (“MEF”) was established to define “Carrier Ethernet” services and standards for telcos and vendors, for access services and for end-to-end services (i.e. across an entire network form access to metro to transport to metro to access). MEF has been very successful in gaining rapid adoption of its standards. It has enjoyed initial success with its highly successful Ethernet over Copper standard, and has built on this success with market offerings implementing the Ethernet over Fibre and Ethernet over SDH definitions of Carrier Ethernet. End result is that major carriers are now offering end-to-end Carrier Ethernet services that guarantee a common language, definitions, and standards and integrity for carriage of Ethernet traffic across an entire network or multiple networks. It provides users confidence and kill set familiarity that the “blue cable” from their router/LAN-network will be presented at their router/LAN at the end destination, and it just works. Importantly, MEF services support Ethernet Internet Access and both Ethernet Private Line and Ethernet Virtual Private Line (E-Line service), point to multipoint (E-Tree service), and multipoint to multipoint VLAN service (known as E-LAN service).
Figure 3 – MEF illustration of breadth of service coverage and transport and Access technologies supported by carrier ethernet (source: The MEF)
The MEF standard for “Ethernet First Mile”, or Ethernet over Copper (“EoCu”) services have transformed the market for Mid-Band Access. Mid-Band is defined by MEF as 2 to 100 Mbps. While E1/ISDN circuits are fixed at 2Mbps, and fibre is usually offered for Fast Ethernet speeds and above (100Mbps), EoCu enables services from 5 to 45 Mbps symmetrical, by logically “bonding” multiple copper pairs (either Unconditioned Local Loop – ULL, or Line Sharing Service – LSS, up to 8 copper pairs with each pair capable of up to 5.7Mbps).
Mid-Band Ethernet services are rapidly displacing E1 circuits, and deferring the cost of implementing fibre.
In Australia, the most widely deployed EFM/EoCu player is AAPT, having deployed MEF compliant Carrier Ethernet over copper switches in over 100 Exchanges to serve the Australian business market, AAPT is setting the pace with many of their clients major 2nd tier carriers and service providers buying their wholesale offering and reselling with value added services.
Carrier Ethernet’s success in the marketplace stems from MEF’s stringent Carrier Ethernet standards that enable ubiquitous, standardized, carrier-class Service and Network to be delivered via multiple equipment and network vendors, and is defined by five attributes that distinguish it from familiar LAN based Ethernet.
The benefits Carrier Ethernet offers are:
- Scales for future bandwidth needs, geographic, applications company expansion AAPT White Paper Series
- Predictability, Risk Reduction, Certification
- MEF 14 provides first performance certification in Communications Industry
- Allows a common world-wide service profile independent of local providers
- Simplified, “Enterprise Style” management puts users in control
- Dynamic, granular, bandwidth-on-demand avoids over-specifying network
- Performance benefits of Layer 2 transport with simplified architecture
- User-driven quality of service allows users to control converged network performance vs. relative application value
- Highest bandwidth speeds available
- Low latency
- Key carrier Ethernet attribute
- Redundant equipment architectures and fast re-routing algorithms
Data Centre & Server Consolidation
- Creates virtual transparent LAN environment reducing risk and cost with highest bandwidth available
- High bandwidth, low latency enables more cost effective use of resources
- Simplifying protocols creates “one-hop” network more suitable to time sensitive protocols
- Allows LAN style management for the WAN
Ethernet Adoption Drivers & Inhibitors
Ethernet at the carrier level can support P2P applications, voice switching and termination arrangements; and a host of consumer applications.
Ethernet adoption drivers for businesses are still very much focussed on the cost effectiveness of bandwidth and the operational efficiencies gained from managing a technology and service that is essentially an extension to the LAN environment. These factors, in combination with the cost benefits of not having to administer and maintain a Layer 3 network where it is not needed (e.g. interconnecting LANs), represents the most compelling drivers for Ethernet services. The continued convergence of voice and data at the Enterprise level and to the Exchange, consolidation of businesses data centres adoption of Software-as-a-Service (SaaS) and the implementation of Cloud Computing are expected to be the major drivers of Ethernet adoption moving forward.
Conversely the main drivers for adoption of Ethernet services are:
For the Service Provider – dramatically lowered cost of implementation; and
For the end Customer – dramatically lower price-per-bandwidth and easier interface and implementation using standard Ethernet IT products.
The Table below provides a cost summary and comparison of commonly used services.
Table 1 – Cost comparison of commonly used business services
However like any technology or service there are a number inhibitors to adopting Ethernet, these include:
- Potential interoperability issues when traversing multiple carrier networks who are not MEF members;
- Potentially difficulties in significantly scaling up networks originally designed for best-effort high-speed internet. However to a large extent addressed by enhanced and advanced VLAN manipulation, VLAN Cross-Connect, QoS and multicasting techniques, combined with hybrid Ethernet-MPLS network design;
- Reliance on application QoS and terminating equipment measures to support end-to-end service performance as applications traverse different networks;
- Potential requirement for additional layer-3 QoS measures; and
- Potentially management of VLANs may require additional IT effort.
Current Ethernet offerings
Carrier Ethernet services can connect multiple geographically disperse sites in a scalable and efficient manner. Ethernet can be a point-to-point or point-to-multipoint service or a multipoint to multipoint service, with bandwidth ranging from 512kbps to 10Gbps.
Ethernet is typically delivered via two architectures
- Multi Edge architecture (involving one VLAN for each service—a “Service- VLAN”); or
- Consolidated, Single Edge architecture (involving a single VLAN per subscriber—a “Subscriber-VLAN” also known as a “Customer-VLAN”).
There has been notable increase towards Multi Edge architecture.
Carrier Ethernet enables Service Providers to offer:
- Point to Point Ethernet Private Line service (EPL);
- a managed Point-to-Multipoint service (E-TREE), for applications to broadcast and telemetry networks; and
- LAN type Multipoint-to-Multipoint service (E-LAN) for multipoint L2 VLANs and transparent LAN Service.
Service bandwidth is typically dedicated, hence not shared with other customers of the carrier. Multiple VLAN is supported to logically separate data traffic streams to satisfy both real time and non-real time requirements of business critical applications and general administrative traffic. Typically, traffic is transported in the carrier’s MPLS backbone.
In delivering Carrier Ethernet services, the physical layer can be either fibre or copper. Telsyte observes that for many Australian businesses, Ethernet over copper remains the selection of choice, primarily due to its cost effectiveness, which caters for the growing demand of bandwidth at incremental costs (scalability to 45Mbps and its value in terms of Mbps/$).
Enterprise cloud computing & Ethernet services
Telsyte defines cloud computing as a highly scalable style of computing where processing of applications and storage are centralised and delivered as a service to end-users via the open Internet or over a private data network.
Cloud computing presents the following benefits to organisations:
Cost savings on capital expenditure – decentralized components replaced by thin clients which are significantly cheaper by comparison, in areas of:
- Cost savings on operating expenditure
- Less maintenance and management required
- Significant savings on power usage by requiring less processing power for the end users
Ease of end-user management – centralised management enables easy configuration of end-user access and simplifies applications version update and maintenance
Enhance workforce mobility – end-user can access information from any terminal and location (subject to security policies)
Although there are limited examples of cloud computing implementations in Australia, many Australian businesses are currently evaluating cloud computing for the future. Telsyte’s research has found that around 40% of Australian mid-market and large enterprises will likely take up cloud computing in the next 3 years, with reduction of operating costs being the main driver for the uptake.
However, Australian businesses have also indicated that security and reliability of telecommunication carriage services are the main concerns of cloud computing that need to be addressed. Telsyte has identified a number of key issues around telecommunication carriage services that need to be considered when evaluating enterprise cloud computing:
- The dramatic increase in bandwidth demand – increased bandwidth capacity will equate to increased expenditure on data services (LAN/WAN);
- The higher level of QoS and CoS required on the network to cater for real-time and business critical data traffic;
- LAN/WAN Network redundancy and availability becomes extremely critical to ensure business continuity - SLAs on with service provider and carrier needs to be monitored and upheld; and
- End-to-end security across the network.
The above factors will likely increase overall telecommunication carriage, and leverage the lower cost per Mbps of Ethernet access and transport service. At the Enterprise level, the inherit scalability of Ethernet, coupled with the ability to implement quality and class of service over standardised Carrier Ethernet services, brings a new level of flexibility and a lowered cost of service proposition for implementing Cloud Computing services.
Related thought leadership
- Australian Broadband and Fixed Telecommunications Market, 2008 Review & 2009 – 2013 Forecast, June 2008 (Publication Number 80587)
- Australian Business Fixed Line Usage and Directions, 2008 End-User Survey, June 2008 (Publication Number 80428)
- Australian DSL Infrastructure Round-up, 2008 Competitive Analysis, September 2008 (Publication Number 80450