May 1, 2015 — The Wireless Infrastructure Show saved the best for the last session. “Mobile Network Virtualization” was perhaps one of the hottest topics on everyone’s lips and one I found absolutely invaluable. Advancements at the semiconductor metal layer have made superchips that are almost a one-stop device with uber-processing power, which can integrate a slew of functionality, making virtual machines of all types a reality.
The panel discussed how virtual networks offer a plethora of benefits. One huge one is scalability, which is probably the single most disruptive development in network management. It enables the network to put virtual base stations just about anywhere. That is the perfect solution to the number one issue – capacity.
These virtual machines can provide capacity on demand, both permanent and for peak demands and to re-map users. This provides better throughput and it balances the entire network as the dynamics of the networks shift. They can offload traffic close to the source, and feed it via backhaul to “big box” processing centers. Other benefits include better cell edge dynamic and the ability to mix and match applications – much like the Windows computers do with applications. And, they simplify expansion to network build outs.
Sounds like the best thing since the Large Hadron Collider and the discovery of the Higgs boson particle. But as much as virtual machines have to offer, they still face some challenges.
The panel brought to light several issues, one of which is that it is new software, something the carriers aren’t fond of beta testing. MNOs are, traditionally, conservative, the presenters noted. They have developed a carrier-grade reliability for present networks and aren’t chomping at the bit to try something new, such as the middleware required to run virtual systems. And, so far there isn’t a lot of understanding of the technology. A pretty steep learning curve is necessary and carriers aren’t anxious to foot the cost and time of the education cycles. Carriers have a lot of KPIs that need to be met. Current networks meet them but virtual networks don’t have much history with them, yet.
In the end, virtual networks will need time to become a reality, according to the panelists. Capacity will be the great equalizer since current hardware solutions will not work much longer as network demands become more and more dynamic and pressures to keep costs under control mount. Network virtualization will evolve and will have to be adopted eventually, according to the consensus. But not until the same QoS as hardware networks can be proven.
That’s all for now…back to the gala…
Ernest Worthman is the editor of Small Cell Magazine.