In a simplistic world, it is easy to view the mobile phone as the PC in your pocket when the truth is the mobile phone is not the PC in your pocket. The path to low cost or free mobile calling to our friends and business colleagues worldwide is not necessarily through a replication of what works on an PC.
Meanwhile carriers have been using VoIP in their backend, transparently to the caller and called party, for about a decade in order to lower their cost of carrying long distance calls between legacy PSTN phones – both landline and wireless. Adoption of wireless VoIP at the end points with the five nines (99.999%) reliability, call quality and scalability of the current circuit-switched TDM wireless networks requires significant infrastructure advances and developments not only at the device level but also at the network level. The wireless VoIP goal remains fraught with obstacles and barriers that are inhibiting its deployment:
- Mobile devices do not have the internal resources of a desktop or laptop PC. Battery life, memory and CPU speeds are all too underperforming at this time to be able to provide a consumer friendly, high quality service. Even the ability to multi-task with multiple applications becomes an issue.
- Robust, scalable network connections are an issue. Current carrier backhaul capacity, especially in North America, is insufficient to support a scalable service at the level required for the demands VoIP would place on data transmission.
- Call quality issues, especially inherent latency and availability of open sockets, still do not make for a consistent call quality.
- Inherent codecs and other support at both the network and device level within the wireless TDM infrastructure address otherwise device resource hogging issues such as echo cancellation and error correction.
- And, most importantly, it does not provide a carrier-friendly business model. Filling the data pipe while attempting to reduce the billable “minutes of use” (MOU) is not appealing to a carrier executive who’s responsible for increasing average revenue per user (ARPU).
So what is the architecture for a carrier friendly system that would allow low cost or free calling to your Skype contacts (or contacts on other VoIP services). The key here is to move the resource intensive processes off the device and network onto a gateway server. It’s implied in my previous post about learning lessons from SS7: Take advantage of the currently deployed robust, scalable wireless voice channel and build the infrastructure to complete the call via the Skype cloud using the Internet as a signaling system. Envisioned by founders Jacob and David Guedalia and working from a 2006 co-marketing Agreement with Skype, iSkoot, has built the necessary clients and infrastructure to realize upon this architecture.
In this architecture, the login, text chat, presence status and call setup information are all transmitted over the wireless data channel to an iSkoot gateway server. The “last mile” for the connection to the mobile device’s voice channel uses the proven, robust TDM wireless voice service to connect to a cell tower. Note that the actual voice connection passes through a SIP Media Gateway. And for each logged in user, there is a Skype client running, on the caller’s behalf, on the iSkoot gateway server.
The process for the user is almost as simple as making a normal call:
- Install the iSkoot client on your mobile smartphone
- Answer questions about your setup, including SkypeID, password and mobile phone number.
- Sign in – it takes a moment or two to download all your Skype contacts with their status
- Select a contact and launch either a voice or chat conversation.
A key to iSkoot’s user friendliness comes from the way it handles a voice call. As with any cell phone call, you simply want to enter a number or select an address book entry, push the “Call” button and have the call completed with no other intervention. When you initiate a Skype call via iSkoot, you follow the same process to complete a call to a Skype or SkypeOut contact. While there is a callback to your mobile phone, it is transparently auto-answered with no user intervention required. iSkoot is available for many smartphones, including Blackberry, Nokia N- and E-series phones, several Sony Ericsson models and Windows Mobile platforms.
But iSkoot has worked actively with Hutcheson-Whampoa’s 3 Service to implement the Skype architecture on a carrier. In December, 2006 3 initially offered an iSkoot-based service as one component of its X-Series service; this allowed iSkoot to develop experience with building out the back end components required to manage Skype clients on their servers in a scalable environment. And, as discussed in previous posts, 3 launched a fully configured Skypephone in the U.K. market last November; its acceptance was at a level such that 3 accelerated the launch of the Skypephone in its other eight markets.
In an interview with recently appointed iSkoot CEO Mark Jacobstein, who sponsored a lunch at eComm 2008, Mark took me through the evolution of iSkoot’s ecosystem that resulted in its availability through either 3 as a carrier or iSkoot’s website for other mobile devices. Simply put, the various clients were the easy part to build; the backend infrastructure was the real and ongoing development challenge:
- Building scalable servers that could manage multiple Skype sessions, not only for placing Skype phone calls but also for handling Skype chat sessions. Currently iSkoot’s servers are located in the Boston area and Israel1.
- Building, with the help of Voxbone, a worldwide network of “local” points of presence such that users can access the iSkoot servers at minimal cost, if any. For instance, from my Rogers service I can call to local servers in major cities across Canada such that only my “local” minutes are charged during a call.
So what are the features of implementing an iSkoot-based solution that appeal to a carrier? From both Mark’s talk at eComm 2008 and my interview:
- Increase MOU and ARPU:
- Drives up minutes-of-use; even though your call is placed to a local point-of-presence the minutes are counted against the “local” component of your carrier’s plan.
- Outside North America, where the caller pays for calls into a mobile phone, it drives additional mobile minutes.
- Reduce operating and marketing costs
- No termination charges for completing a call to a Skype contact. Termination charges are the carriers’ third largest cost item after network infrastructure and personnel. This is why 3 can offer Skype calling for effectively no additional charge to the end user beyond their monthly commitment for outbound calls to the PSTN (landline or mobile).
- Low customer acquisition costs. Whereas most carriers need to subsidize new customer acquisition through subsidies for the mobile phone device itself driving hard customer acquisition costs into the range of $150 to $300, hard customer acquisition costs for a Skypephone customer, including the cost of a handset, are much less than $150 and no subsidy is required.
It should be noted that while iSkoot solves the international long distance calling problem, solving the roaming problem requires a multi-carrier SIM chip such as those provided by MaxRoam. For instance, when I go to the U.S. and make an iSkoot call I am charged the $0.95 per minute roaming charge independent of whether I make a long distance or local call in the U.S. I hope to test out the MaxRoam service on my next trip to the U.S.
Yesterday iSkoot announced an extended partnership with Skype along with additional carrier relationships where they expect to implement the Skypephone in countries where 3 is not available including Argentina, Greece, Israel, Norway, South Africa, Switzerland and Turkey.
Where does iSkoot get its revenue? As Mark mentioned at his eComm 2008 talk, they and Skype each receive a royalty for the monthly customer subscriptions to 3. In addition, of course, iSkoot receives revenue associated with the phone hardware.
And, if you’ve got a supported handset you can start benefiting from iSkoot’s services directly. I have been using it on my Blackberry 8820 for the past two weeks with good success (especially when I am away from my office) with three caveats:
- Skype needs to provide additional presence and chat support such that dual logins (PC and iSkoot) provide consistent results in terms of where a chat message will go.
- iSkoot needs to ensure that closing a chat session on the mobile device also closes it on their server such that the session does not return to the device. At the moment the only solution for this issue is to actually Sign Off (to shut down your chat sessions on the iSkoot server) and sign back in again — at least once a day.
- While you can readily make SkypeOut calls to SkypeOut contacts in your Skype contacts, there needs to be integration into the device’s native address book to allow one step calling of your native address book’s contacts. (This is already a feature on the Skypephone.)
If you don’t have a Skype phone you can sign up for the iSkoot experience here (supported devices are listed). Keep in mind, while it appears to work quite satisfactorily, it is still in beta trials so your feedback is appreciated.
Mark raised many additional points in his interview; I will bring them out in a future post. And, finally, no there is no iPhone support at the moment; even doing an iPhone client is still under investigation, given all the constraints associated with the iphone’s SDK release.
1 Note that 3 has positioned servers within their own Network Operations Centers to support their iSkoot-based Skypehone activity. Subscribers can access them at no charge via any 3 cell tower.
Tags: iSkoot, Skype, Skypephone, 3 Skypephone, Mark Jacobstein, Blackberry, Nokia N-Series, Nokia E-Series, Sony Ericsson
Powered by Qumana
Related articles by Zemanta
- Startup lessons – iSkoot style (startupcfo.ca)
- Mobile VoIP Startups Looking Beyond Cheap Calls (gigaom.com)