Next Gen Wireless – Twitter, Foursquare and Spurring Innovation
Since the launch of the iPhone, carriers, and in particular AT&T, have been in the news a lot. Most of the press is around why the service providers fail at providing a decent user experience. From dropped calls to dead pockets, people have come to accept what would otherwise be a totally unacceptable SLA (service level agreement). Which is weird because people use their cell phones for business a lot these days (certainly any roaming professional or salesperson does almost 100%), and so these aren’t just dropped calls to our friends and family.
The next generation wireless standard called 4G, often referred to as LTE (long term evolution) is finally around the corner. LTE is actually a subset of EPC (evolved packet core), which is basically 4G. Confused? Anyway it’s coming – so what new usage paradigms will this new wireless network bring? Will our ‘always on’ connections be more reliable? What about voice and dropped calls? What about battery life on devices like the iPad and iPhone, when they eventually get 4G / LTE (won’t be 2010, but will surely be in 2011).
Here are some thoughts:
1)Let’s think about the carriers and ‘unlimited’ data for a minute. How did this happen? Pre-iPhone, network congestion / loading was not a problem, but consider that it took the wireless industry over 20 YEARS to reach a point where they offered unlimited voice (flat rate billing), and they offered unlimited data almost right out of the shoot for data intensive smart phones. Interesting! There are practical limits to voice usage, but there aren’t really with respect to data. Flat rate billing is definitely spawning a different type of user behavior.
2)Important trend – the decoupling of revenue and data: the relationship between traffic and revenue is diverging and is no longer 1-1. Traffic is increasing exponentially and revenue asymptotically. Carriers need to expand services and satisfy customer expectations to remain competitive.
3)LTE is the first ‘all IP‘ wireless networking standard. This has a couple of implications. First off, there is no inherent provisioning for voice (in the legacy circuit-switched sort of way). Instead, all voice is packetized and sent as IP (internet protocol) packets. This means LTE will be an entirely new way of handling voice calls, hopefully one in which the underlying layers in the IP stack can manage better reliability.
4)The ‘all IP’ internals of LTE really lend themselves to providing a more seamless network integration above today’s hodgepodge of standards. LTE has session intelligence suck as location info, time of day, access type, subscriber type handover status, device type etc. These can enable redirection of traffic based on higher layer data. Operators could use the session intelligence to insert data or to improve the user experience.
5)Location based services are the rage today FourSquare, Yelp etc. Interestingly, polling-based applications like these applications put a tremendous strain on the network. They don’t suck up large amounts of bandwidth, but they are signaling hogs – i.e. they are constantly polling with the network for location and status updates. Many people don’t realize that this massive increase in signaling has strained today’s networks tremendously. LTE does a much better job of smoothing out network bottlenecks…
6)But…Will there be best practices and guidelines in regard to how these interactions should occur? Application comparisons across devices demonstrates that the same app can be exercising the network differently. From a performance and efficiency issue (can look at a cost distribution) these forces need to be measured appropriately based on network conditions. For example, it’s expensive to the network to constantly poll, as is done in Twitter applications. This is an expense at the signaling layer, not in bandwidth! How do we map the cost distribution to reliability and user experience?
7)So.. Congestion is highly device and application dependent today. Carriers must get out of the silo’d network approach. Flow control and congestion control need to be used in a ‘surgical’ way for wireless instead of the ‘broad brush’ global way which has been adequate for wireline. E.g. they need to take real time data from the network, correlate it and manage it in RT, then also touch the device with over the air provisioning (feedback loop). But do the carriers and network providers a) care enough b) really understand the trends?
8)At millions sold, the Kindle was a breakthrough for fixed / integrated embedded data-cards, and this ‘always on’ device format will be happening A LOT more in future (digital signage, home wireless enabled devices, kiosks, wireless vending machines etc). LTE offers better service for the types of transmissions properties which these systems tend to need. Specifically – chatty, secure, applications with a lot of heartbeats, back and forth communication etc. for example an inventory control or surveillance application. But without the type of provisioning and smart network architecture these advantages will be normalized.
9)Net neutralityVery interesting debate bottom line is operators have a responsibility to manage their network to optimize QoE and profitability. If they do this in a conscientious way they *should* treat the needs of the user importantly and find the right balance. Instead of ‘controlling an application’ they must make policy more about managing the network in a way where unused BW is intelligently allocated ‘getting around congestion points etc. In-home Femto-cells are self organizing and self optimizing’ this is an example of a product today that offers reduced complexity and enables consumers.
This is not an exhaustive list but serves as some groundwork for how next gen wireless must take into account more about the user and application in order to succeed. The carriers are testing and deploying equipment from the telecom network providers today (Alcatel, Ericsson etc). Only time will tell if the complaints emanating from the AT&T / iPhone press will really be taken into account, and whether the cost dynamics and past deficiencies in the industry will allow the carriers make the right decisions? let’s hope.
Related articles by Zemanta
- Verizon: Where a Megabyte Costs Almost as Much as a Stamp (gigaom.com)
- What if the iPhone Cheated on AT&T? [IPhone] (gizmodo.com)
- AT&T boosts network investment (news.cnet.com)
- Analysts: Apple Tablet not necessarily on Verizon Wireless (macworld.com)
Zemanta helped me add links & pictures to this email. It can do it for you too.