Enterprise-Grade Wireless Applications: just a gadget?

ACM Queue May issue features a survey article about wireless technologies in enterprise. Here, enterprise wireless application is defined as:

"...extend enterprise applications to wireless devices used by employees and clients"

I prefer the notion of "Mobile Enterprise Applications" for the same thing. Note that this does not encompass the wireless network infrastructure such as WiFi or WiMax or whatsoever; we are talking about running enterprise applications on handheld/cell phones to improve operational productivity and reduce cost. Some ISVs already begin to ship mobile CRM or ERP software. For example, iEnterprises Inc. has a CRM software running on RIM BlueBerry.

The article identifies the following challenges of developing, running, and supporting wireless enterprise applications:

• Devices; still no fit-all PDA. The market is segmented into consumer-centric feature-rich PDA and enterprise-centric PDA that is predominantly used for work as an extension to an enterprise application.
  
• Networking; for most applications, wireless data rate is not the problem; CPU and battery are the keys.
• Software; trend is the rise of monolithic system that integrates many otherwise separated systems. Well defined user interface and machine communication interface (XML and SOAP) are essential to mobile enterprise applications as well.
  
• Support nightmare
  
• Security; data security on the device (encryption?), on the server (Paris Hilton's T-Mobile story), and in transit (VPN, SSL, WiFi security, etc). Also, the solution must be easy to use.
  
• Uniformity; determine a standard set of requirements for evaluating offerings. "One such uniformity requirement that merits special mention is the partially offline scenario", which dictates that the wireless platform must insulate connection-oriented applications as seamlessly as possible from network interruption by emulating an always on and highly reliable network. Asynchronous connection and data push are needed.
  
• Manageability; company policing the usage of devices

Implementation issues include choosing the software model (the author suggests thin-client over browser based), choosing the programming platform (most likely a VM such as J2ME, Brew, or .Net Compact Framework), data caching, etc.

One thing worth noting is the author claims that
"J2ME and .Net CF share a common weakness,..., in that they do not assist developers with the "hard" problems of wireless development, such as solving the offline/partially offline issue, providing secure network transport, or integrating asynchronous push technology."

"J2ME and .Net CF share a common weakness,..., in that they do not assist developers with the "hard" problems of wireless development, such as solving the offline/partially offline issue, providing secure network transport, or integrating asynchronous push technology."

Is this true? As far as I know, .Net CF provides SqlDB to manage local data store, and has some classes for synchronization when the device is online. Don't know about J2ME in this regard.

Mobile Media Making a Reality

This month's ACM Queue is centered at "Going Mobile", coincident with IEEE Pervasive Computing magazine. One of the stories, "Mobile Media: Making it a Reality" is worth reading. It talks about two research projects at HP labs: one is InHand, a personal context aware mobile services that leverage RFID technology; the other is a Mobile Gaming Service Platform.

It you look at the mobile networking side of Inhand, it is nothing new: like a typical Near Field Communication system, users will be able to use a mobile device equipped with an RFID reader to access information provided by some tagged objects in proximity. The device talks to a backend server via cellular connection or WiFi.

Application scenarios include move goers swing a cell phone to a movie poster to watch the trailer, pre-order ticket; shoppers at a grocery store point a cell phone to food for a recipe and check out automatically. Or, when you are shopping at B&N, you can use your cell phone to read the ISBN code and send an SMS to a system that will return the price of the book at Amazon.com.

What distinguishes InHand from other similar projects is its context-awareness: user's information access experience is largely personalized based on her profile and other parameters such as networking conditions, device form-factor, etc. You can of course imagine this contextual information may involve advertisement. These all happen on the backend server that services the request from a cell phone that combines RFID and user information. InHand sounds like a technically workable idea, and I guess consumers would like to try this if available. But the difficult part is the business model: Assuming there is an independent content provider servicing the media, how much does this benefit the site(movie theatre, for example) that provides such facility? If the connection between the mobile device and the media server is cellular, will a user pay for this?

Multiplayer mobile gaming is getting a lot of attention recently. As the author suggested, "multiplayer games, which promote community communications, tend to engender gaming loyalty". The MSPP platform is based on the IMS (IP Multimedia Subsystem) of 3GPP, which is covered in my book "Smart Phone and New-Generation Mobile Computing". Two major challenges exist: transport layer requirements for mobile gaming where TCP and UDP do work well, and gaming service user interface design in association with presence service and person-to-person communication using real/virtual identify in real/gaming world. In addition, end to end security can only be achieved if the media proxy who is responsible for transcoding (adapting streams for client capabilities and time varying network conditions) do not need to decrypt the steam in order to transcode. The researchers at HP labs have proposed a scheme called "Secure Transcoding" for this purpose.

There is no doubt that mobile gaming could be a kill app for 3G data services-just look at online gaming.; but it still stays at the vision phase-no such services are available now-voice is still No1. Cell phone's tiny screen is certainly a problem, which makes me think about putting text-based games such as Mud in the first place when it comes to mobile gaming.

On the other hand, group gaming, i.e., players using WiFi or Bluetooth to play in a physically proximate setting may get some traction, especially when you consider Sony PSP and Gameboy DS's networking capability in terms of ad hoc communication.

IEEE Pervasive Computing on Smart Phone

The April-June issue of IEEE Pervasive Computing has several informative articles on smart phone application, exemplifying some exciting advancement in smart phone research and development. The following is a quick briefing of the issue.

Swiss Army Knife or Wallet?
As always, the editor in chief, M. Satyanarayanan, brings the readers an insightful perspective into smart phone computing. He points out that

What is the most valuable functionality to cram into a small, lightweight device that has a long battery life? Every designer of mobile computing hardware wrestles with this question. A large reward awaits anyone who can produce a good answer and follow through with a good implementation. The iPod's success in displacing older portable music devices shows the power of combining the right device functionality with a good form factor and user interface.

Apparently, due to the restriction of the form factor, it is impossible to offer everything with a smart phone. Smart phone must be a phone in the first place - everyone (well, almost) agrees with this. Beyond that, fancy features are being added to make the phone sexier to teenagers as well as business professionals. Camera phone, email, gaming, IM (think about T-Mobile sidekick), mobile TV, ..., you name it. The design philosophy is named "swiss army knife" by Mr. Satyanarayanan: cram many things into one, but each one only provides largely reduced/constrained functionality. The article adds that instead of designing swiss army knife like mobile devices, it would be more viable to have "wallet" like smart phones which are a must-have in everyday life and carry in-direct functionality (credit card, cash, etc) that leverages the surrounding environment. This is indeed a good point-rather than fighting with limited computing, communication, and storage capability of a mobile device, try every possible means to make it easily fit into a computing environment (computers, the Internet, wireless links per se). Again, user interface is a key.

Mobile Hardware
Japan's T-Engine platform provides boards and middleware for a wide range of mobile devices, ranging from smart phone to home appliances to wireless sensors/switches. Stamp-size LCD display hanging close to the eye has long been marketed as next generation display technologies but only managed to create niche market. However, near-eye display with VGA or high resolution may have a good chance to replace the tiny cell phone screen (QCIF or CIF), as long as the cost is reduced to a level around one or two hundred US dollars.

New Products
A family location-tracking application provided by uLocate sounds like a clone to the family car tracking application-both use GPS, and parents can view the locations of their kids over the web. Phone-controlled video camera like Sony Ericsson's Mobile Cam ROB-1 are nothing new-I still remember Nokia Remote Camera. But the ROB-1 is a mobilizable camera-it can move! Kodak's Wifi-enabled digital camera may enable a new fad "DC-blogging", in addition to using camera phones to upload pictures to blogs. 2G micro-drive from Toshiba further sheds some light on integrating video related applications into a smart phone.

The Smart Phone, A First platform for Pervasive Computing
The article identifies two kill apps for ubiquitous computing: person-to-person voice communication and SMS (over 100 million text messages are exchanged daily in China). No one will argue. Next one? no one knows for sure. But it is very likely that smart phone will become the "first" platform for pervasive computing, facilitating novel applications such as mobile social networking services (remember dodgeball?), camera phone enabled phone/video blogging, personalized information sharing(p2p?), and of course anything about location based services.

Enabling Pervasive Computing with Smart Phones
Using smart phone as an end point for an information utility or service, a number of services can be devised, including spatial navigational assistance (mapping) along with local search, shopping assistance, entertainment, healthcare monitoring and diagnosis. Examples of projects are mXpress(conference/event assistance), MyGrocer (RFID-assisted shopping), E-Care(mobile data collection, monitoring, alerting, etc, in healthcare), and TellMaris(3-D mapping). The article shares some of the lessons learned from design of those systems. The article continues to argue that using mobile phone as a remote controller "has appeal in the at the mobile phone fits this role well", and will "become a manifestation of the pervasive service" as a point of contact. A mobile phone is of course a network hub to provide wireless connectivity for connecting equipment, just like a model, so we don't need to worry about the form-factor. Mobile phones as ID tokens holding credentials extend its functionality to another level: mobile payment, mobile banking, mobile bidding, etc. The article concludes that "Until practical ubiquitous interfaces and services emerge that allow for a greater wealth of interactions, the smart phone will remain the best available personal computing device". Indeed, smart phone itself may disappear into the environment in the future; but for now, it is the best candidate of universal mobile terminal for a large array of emerging applications.

Social Serendipity: Mobilizing Social Software
Serendipity is a Bluetooth based mobile socialware that facilitates interactions between physically proximate people sharing similar interest. It can be seen as an extension to "online introduction"/social networking applications such as Friendster and Match.com. A J2ME MIDP application on a user's Bluetooth phone periodically scans other Bluetooth devices in proximity (assuming many such devices are in 'discoverable' model). The system may use employ dedicated device called BlueDar to do the same thing, then uses WiFi to transmit detected BTIDs to a serendipity server, which performs profile matching as well as privacy driven SMS notification. Relationship inference relies on some information retrieval methods that basically computes the level of similarity of two sets of pieces of information-just like in Peer to Peer networks, we match user's interest with their query strings. The article has a list of mobile social software, including Lovegety, Gaydar, Cell tower/SMS friend finder(Dodgeball), Experience Ubicomp Project, Social Net (using specialized wireless computer), Hummingbird, and Jabberwocky.

Supporting Social Interaction with Smart Phones
7 experimental smart phone applications developed by researchers at University of Birmingham exemplify smart phone's ability to support a wide range of social interactions, which would be difficult if not impossible to achieve without the smart phone technology. The author argues that in a pervasive environment, phones exist in a social setting where the focus is communication, not computation. Smart phones's computation capability should be leveraged to augment communication. The article uses a simple shorthand notation for interactions between people: 1-1 (person to person), 1-N (person to group), N-N(within group communication), 1-all (person to world) and N-all(group to world). A smart phone application is useful only if it supports more than one type of communications. The systems introduced in this article are: Blue dating (locally stored personal information compared to Serendipity's centralized approach), Community building such as Joke sharing (Chinese SMS users should be very familiar with this app) File Sharing (mobile P2P), Public Space and Share Space (like a mobile BBS but without an admin to constantly monitor and manage posts), Intelligent Multimedia Messaging System (placing ipaqs or smart phone on the office doors and remotely control the display while allowing visitors to use it), and Mobile Blogging (Bluetooth based picture posting via HTTP)

WiMax: State of the Art

As expected, Intel Introduces New WiMAX Silicon Solution To Expand The Reach Of Broadband Internet Access on April 18, 2005. Its rival, Wavesat who claimed to make the world's first WiMax chips, started to ship their WiMax chips in January. Both companies have partnered with wireless system manufacturers to market WiMax based products.

Quote from Intel's WiMax press release:

Previously codenamed "Rosedale," the Intel® PRO/Wireless 5116 broadband interface device is based on the IEEE 802.16-2004 standard, giving carriers and end-users the confidence that equipment from different vendors will work together. WiMAX solutions based on 802.16-2004 enable the creation of high-speed, fixed wireless broadband networks, providing Internet connectivity, Internet Protocol (IP) and TDM Voice capabilities and IP-based real-time video at high speeds."

Note: the 802.16-2004 is the "fixed" WiMax, whereas 802.16e is the mobilized WiMax. dailywireless.org reports thatMobilized WiMax Solutions from Adaptix, SR Telecom, nex-G, Wi-LAN's Mobilis and Proxim Pre-WiMax are also being promoted - "pre-802.16e" or not.

Quote from DailyWireless.org:

While many vendors have been releasing "pre-WiMax" or "WiMax-like" gear based on their own proprietary technology, no WiMax equipment has yet been certified. The WiMAX Forum will initially certify equipment based on two profiles: TDD and FDD in the 3.5 GHz frequency band with 3.5 MHz channelization. Based on market demand and vendor product submissions, more profiles will be added in 2006. The WiMAX Forum also has committed that all future enhancements to the baseline profiles will support backward compatibility. To demonstrate their commitment to these timelines, several equipment OEMs demonstrated working equipment at the member meeting.

The following figure shows spectrum usage of Wimax.

Pre-WiMax on rails

T-mobile starts to offer free WiFi access for rail commuters between London and Brighton using a Pre WiMax technology. When the WiMax signal gets lost, the system migrates to GPRS. Here is the big picture: a commuter's WiFi interface talks to an antenna mounted on the roof of the carriage, which in turns talks to base stations by the side of the track. From there an ADSL 2Mbps link is available to an ISP.

The system is actually using an 802.16d standard, rather than the highly anticipated 802.16e. That is why it is called Pre-WiMax.