Wanted: Mobile Applications Acceleration

www.stampede.com they already have a windows mobile client and have had it for some time

Let me try adding some spices to the dish Larry just brought up:

1. Let's first clarify the subject "mobile" a bit further. The category includes WiFi LAN, WiFi Mesh, the up-coming WiMax, then of course, the 2G, 2.5G and 3G networks. Communication terminals or devices in these networks are all mobile. Acceleration technology requirements for the different environments can be vastly different;

2. Secondly, requirements for acceleration from the two spheres of the market, namely the consumer and enterprise, are very different. Applications and therefore their appetite for bandwidth differ substantially;

3. Third, hardware platforms and operating systems, upon and under which applications run can be as far apart as apples and oranges. The resources available in these environments for the implementation of acceleration vary greatly. For example, try comparing Windows Vista+Mobile PC to Linux+consumer mobile phone (not Smartphone);

4. The term "Acceleration" is an umbrella that covers anything ranging from basic TCP acceleration, UDP header compression, In-tunnel transport protocol change, flash compression, latency mitigation to WAFS and CIFS acceleration, content caching, and more. Implementing all into one single box of acceleration gear proves to be unrealistic even for giants like Cisco, let along onto a mobile phone. For example, try imagining implementing Riverbed's success of WAFS technology into a cell phone.

With the landscape in view, we are ready to delve into the challenge Larry brought up. And let me focus on addressing the issue in a generic way, and perhaps what the thread began with and intended for - acceleration on a mobile Smartphone and on 2.5G or 3G network.

With physical limitations on bandwidth in the 2/2.5G data networks application developers thus far built the mobile apps with the constraint clearly in mind, so the volume of data moved by these apps across networks is small. Pressure for acceleration has not been great until now.

Mobile operators have traditionally been accustomed to voice services and in general insensitive to the looming data challenges just below the horizon. In addition, they were simply too busy to deal with the explosive growth of voice service subscriptions. There has been no bandwidth to worry data, Déjà vu of the two telecom giants missing the Internet boom, anyone?

Acceleration technologies from the vendors in this space of the acceleration market today almost exclusively call for the deployment of acceleration or compression equipment, terminals, soft agents symmetrically, i.e. deployment at both ends of a communication path. While this is logical, sensible, feasible and practical in an enterprise environment (private leased circuit), they are pale facing the public Internet at large and the wireless public network, where traffic goes anywhere and everywhere. Tunnels of which, most, if not all, acceleration techniques rely, can not establish. Asymmetric acceleration technology is a must.

Mobility solutions provided today by various vendors such as Packeteer, Riverbed, Bluecoat, Citrix and the like tap resources from the powerful PC hardware so their boundary is within the perimeter of the enterprise and campus networks. Solutions for the mobile Smartphone must function under a very tight budget of CPU and memory resources. Code size does matter. Consumption of CPU cycles by compression, memory occupation by acceleration, discs, and battery life are all prohibitive barriers. (just saw Stampede's enterprise solution as exceptional, but still needs two-way)

Existing vendors also face a tough challenge of adaptation to a strange world, where Symbian S60, WM 6, iPhone OS, Android and Linux rule.

One can go on for another page of reasoning, but I hope this explains the situation we face.

However, solution may arrive on the scene soon. Stay tuned. :-)

Wopro