In Amsterdam, a ‘Cloudy’ Forecast for Broadband
By Uwe Lambrette, Director, Cisco Internet Business Solutions Group
Amsterdam may be the one place on earth where it rains more than it does in London. So, it was no surprise that I encountered stormy weather on my flight to Broadband World Forum (BBWF) 2012. As things turned out, the conference theme and the weather were clearly aligned, since the BBWF is fiercely embracing evolution to cloud. Here are some core themes that emerged as I shared some of Cisco IBSG’s findings at the conference:
Cloud 2.0: Most service providers (SPs) have already launched an initial cloud offering and are now beginning to measure scaling and growth. The initial offering is often a stand-alone cloud solution, typically focused on infrastructure as a service (IaaS). Once their original implementation goes live, SPs often need to focus on the following improvements:
- Building a platform that aligns with common customer use cases (e.g., hybrid cloud). SPs don’t usually start with hybrid cloud. Much like enterprise, they primarily build their own (siloed) solutions. As a result, the “cloud silo” in the SP is often not aligned with or targeted to an enterprise that has its own partial solution in place.
- Scaling and go-to-market strategy. Once the technical launch issues are ironed out, the next area of focus is go-to-market. SPs have to build out an integrated offering (e.g., resale channels for SMBs) and embrace a professional-services approach.
- Improving the economics. SPs are using open-source software like the cloud.com stack. Korea Telecom is but one SP taking this approach. This is in line with Cisco IBSG’s findings in enterprise, where companies with a high level of expertise in cloud increasingly move to open-source environments.
Cloud Service Brokers (CSBs): Service brokers were another core theme, as they are a critical control point for SPs. A CSB platform enables the SP to retain customer ownership even when individual offers come from various providers, whether hosted in the SP data center or elsewhere. (For our purposes, the CSBs being discussed are those bundling diverse offers for a single customer target segment. Other CSBs focus on bundling single offers for a diverse set of customers. Enomaly, with its IaaS offer, is an example.) All SPs have launched CSBs. But is the model sustainable? In the long run, newer technologies like HTML5 could enable customers to pick SaaS vendors easily and directly without paying an intermediary’s distribution margin. The CSB must truly add value; having probed one CSB, my sense is that sustainable value is not easily defined. Here are four observations on the potential success of CSBs:
- On the downside, the critical issue of data integration (enabling SaaS applications to share and exchange data) is still “being worked out.” Without data integration, it becomes difficult for CSB customers to ensure data consistency across a more extensive set of SaaS applications. SaaS “sprawl” threatens the integrity of corporate IT. With data integration, CSB survival is secured.
- On the “so-so” level, a CSB is attractive only when it enables a broad ecosystem. An application developer, for example, does not want to customize software for a multitude of CSBs, as it is too resource- and time-intensive. The CSB must therefore command a wide range of ecosystem partners to benefit from the network and scale effects, and also to be in a prime position with developers. It is essential that the integration of the software offering into the CSB framework is simple enough that the CSB does not become the bottleneck.
- CSBs will not “cut it” in the SMB market, since SMBs want additional, trusted personal services. Thus, SPs are advised not to rely on portals alone when selling cloud services.
- A CSB’s key benefit could be enabling a commercial framework in an existing distribution chain, so that resellers and channel partners have clear and reasonable incentives to promote.
- Cisco IBSG has been told that enterprises are not interested in buying from “middle men” (like CSBs) as they just add margin and eliminate customizability. In my opinion, the longer-term survival story for CSBs has not yet been decided – many SPs are experimenting: SingTel, Telstra, Telefónica, Deutsche Telekom, KPN, to name a few.
Network and Cloud Integration: This is a key theme for SPs, and high hopes are placed on this vision: Higher agility, custom service offerings for enterprises, integrated SLAs, lower costs and lower complexity.
- For IT Services. Whilst this is not only about delivering IT services, it heavily plays to delivering IT services. My concern is though that many workloads don’t require this integration; for SPs, winning in this market is more about extensive feature sets and ease of buying experience (bypassing corporate processes like partisans). Those pursuing network and cloud integration forgo quick-win cloud workloads. Where high performance is required, network and cloud integration is indispensable, as are hybrid cloud offers. This is no easy game, and SPs who want to win here must have high competencies in their offers from the application to network level, including hybrid cloud. Further, an agile network-cloud combination requires significant OSS efforts. OSS is often the singular area causing most trouble in time to market and service delivery, on the other hand, as highlighted in some keynotes. So this proposition poses additional strains on an already stressed asset. Delivering against this vision will be possible only with partners, such as systems integrators. The only immediately appealing SP route appears to be horizontally scalable, quality-dependent workloads, including collaboration and video. These have the added benefit of natural proximity to the SP’s core business.
- SDN Hype: The “cloudification” effect (“let’s use commodity hardware and move all complex networking operations into a cloud orchestrated by low-cost, open-source software”) is also reflected in discussions of software-defined networking (SDN). In my humble opinion, SDN will offer limited appeal in the broader SP context. In the era of the Stuxnet worm, there is too much concern about customers and perhaps employees interacting in “undesirable” ways, increasing the liability of SPs. It is already enough of a challenge to manage all policies for network configuration; it is quite another to give thousands of business customers access to a configurable network while providing them with competitive and reasonable flexibility.
Deutsche Telekom, for example, will launch Terastream, where some of the higher layer network functionality such as CDN and IMS will be integrated in a cloud service center. It’s not a proper SDN yet, while trials to use it in the RAN context are about to begin. In IBSG (my team) we think any SDN agility benefits rely heavily on the business architecture, and how this is matched with OSS. OSS is the Achilles heel.
Nonetheless, SDN will find its application in focus areas such as machine-to machine (M2M). In a broader context, SDN complements virtual data centers and the integration of a complex set of resources in a virtual container as part of a multi-tenant cloud. In this case, the resources themselves are entirely virtual, and they are not shared among multiple users even though they execute on a multi-tenant cloud platform.
Cynics have told me “SDN is just a pressure tool to get better deals.” They believe that most suppliers—when presented with a ferocious threat to deploy SDN—will knock a few percentage points off their, say, maintenance contracts, rather than risk losing a franchise.
In summary, I think SDN will be very effective in well-defined segments of the market, but it will likely stay out of the wide area of “multi-tenant” service provider networks for the time being.
At the Broadband World Forum, cloud was front and center in nearly all discussions, and I believe this presages many further positive developments. It may have been dark and stormy in Amsterdam, but there was no shortage of flowers throughout the city. Similarly, a thousand flowers will blossom under cloud, and a world of many clouds will continue to evolve.