Jarich: Why the mobile core is more important than ever
Compared with advances in mobile devices, M2M business models, app stores or application development, the mobile core doesn't generally get the attention reserved for other parts of the mobile ecosystem. As we begin closing out 2011 and looking forward to 2012, however, I've been putting a lot of thought into this space. Why? It all revolves around an expectation of how mobility is evolving--a vision that directly plays to operator buying criteria in the mobile core, the way vendors are selling their core solutions, and how need to evolve their solutions.
There is no shortage of distinct visions about the future of mobility. Most of them, however, tend to revolve around a number of agreed upon (or, at least, hoped for) expectations.
- Lots of Data. Any future mobility story begins with the concept of the "data deluge"--more and more users, generating more and more sessions, each consuming more and more data.
- Lots of Applications. If we left out anything above, it's the role of applications. At a basic level, compelling mobile applications are responsible for skyrocketing data demands.
- Lots of Business Models. Consider LightSquared. It's gotten a lot of press and visibility, in part, because it's a new model for mobile broadband. Add in operators looking to drive revenues from their subscribers as well as the over the top (OTT) content providers driving traffic across their networks or carriers simply looking to monetize differentiated QoS capabilities.
- Lots of Places & Lots of Cells. If operators are looking to drive mobility deeper into the lives of their subscribers, network ubiquity moves from being an interesting concept to an imperative. This means coverage along with the targeted capacity to make that coverage meaningful. This is why small cells have taken center stage in terms of operator network architecture investigations.
- Lots of Technologies. Even as LTE networks are getting rolled out, there is a realization that 3G networks (and 2G in some instances) will live on for years. What's more, Wi-Fi will join them as a solution for traffic offload and/or international data roaming.
- Lots of Missions (many critical). By definition, some mobile broadband applications will be mission critical. Others might be considered mission critical on a de facto basis--branded or enterprise applications where an operator's compensation depends on a solid user experience.
The use cases bandied about at conferences and in vendor messaging all speak to these expectations and network directions--this vision. M2M communications where anything that can benefit from connectivity is attached to the network. Mobile commerce where the phone becomes a wallet, ID and set of household keys. Context and location-aware communications. Ubiquitous connectivity. Each requires a network supporting user and data scale, location scale, multiple applications and technologies...and does it with the reliability of a mission critical network. The foundation of this network is the mobile packet core, the conduit from their RAN to the service layer and broader data networks.
Ultimately, operator visions and business model demands get translated into buying criteria--principles that guide whether or not a specific solution meets their demands. These buying criteria impact nearly every component in the network from backhaul links to BSS/OSS platforms to base stations and the base station shelters that sit at the bottom of towers. In the mobile packet core, the link between the service visions outlined earlier and operator buying criteria rests on four principles.
- Scalability. The notion of "scale" gets raised often in telecom product comparisons and performance claims. To be sure, the need for scale in the packet core is driven by capacity evolutions in the device, radio access and backhaul layers. Beyond that, however, the reality is that scale in the packet core must be measured along a number of different axes: throughput, switch fabric, RAN fan-out, simultaneously attached users, IP bearers or PDP contexts, service data flows, transactions per second, policers...the list could go on forever. Just as important is how user plane vs. signaling traffic scale independently (or not).
- Form Factor Diversity. Going forward, however, there's a need to site mobile data gateways closer to users. Why? With the gateway closer to the subscriber, service latency can be improved. At the same time, traffic can be offloaded more immediately, economizing on transport costs. Perhaps most importantly, small cell RAN architectures argue for localized packet core functionality.
- Service and Innovation Flexibility. Along with "scalability," two of the most used terms in the mobile packet core lexicon are "innovation" and "flexibility." This overuse can be blamed on their importance and visions of integrating network insights with subscriber knowledge in order to deliver monetized, context-aware services. The goal of making the packet core an engine for innovation across multiple RAN technologies implies a number of different things. Support for standard interfaces to billing systems, policy engines, IMS and service delivery platforms along with critical mobile data applications and supports (security, content filtering, analytics, content optimization, content caching) and the flexibility to support the applications we don't yet know about. Support for technology agnosticism which reaches from 2G to 4G to Wi-Fi. Support for API exposure as a way for operators to fully monetize their network assets.
- Reliability. Reliability--aka availability or uptime--belongs at the end of this list. Not because it's the least important of the criteria, but because it's probably the least refutable based on the number of subscribers and diverse network elements a given gateway touches.
Comparing these buying criteria against vendor messaging, however, suggests a gap. As LTE and 3G solutions like HSPA+ have gotten rolled out, scalability has been a driving focus for many vendors. Throughput and user capacity have been front and center in marketing. New form factors have been rolled out to support new architectures thanks to low-end scalability. Most major vendors have touted their work with performance testing companies to prove out their scalability claims. Messaging around service integration and application flexibility, however, has been largely missing. It's implied in Alcatel-Lucent's Application enablement story or NSN's focus on Customer Experience Management. We see examples in Cisco's MOVE launch or, more explicitly, in Juniper's MobileNext portfolio and its Junos SDK.
If operators hope to execute on their full potential of mobility--and vendors hope to help them with that--we need to see a greater focus on more than just core network scalability. Vendors who've yet to fully outline their strategies and solutions need to do so. Vendors with a leg-up on this front, need to prove the value based on real-world deployments and application partners.
Peter Jarich is the Service Director leading Current Analysis telecom infrastructure practice. Follow him on Twitter: @pnjarich.