5. NOVEL SOLUTIONS BY LEVERAGING MOBILE PLATFORM

In Area 4.2, we mentioned how we can use known techniques to protect against sensor-sniffing strikes. Each strategy satisfies only some of the suitable qualities in Area 4.1. This is no surprise, since we face similar difficulties as the common issue of viruses recognition.

However, there are important differences between sensorsniffing strikes and common viruses strikes. With receptors on Lenovo A916 Cellular systems, for example, when the customer cannot decide whether a system is reliable, he can let the system inform him if the system accesses the mic (so that he will not discuss delicate details when the notification is on). By comparison, this allow but inform strategy is unsuitable for protecting against common viruses strikes because once the system allows the viruses to read a file, the confidentiality of the file may be breached instantly. The exclusive factors of Pandawillhttp://www.pandawill.com/ Cellular systems also offer us with possibilities to examine novel immune system. We offer three such illustrations.

Context-aware system profiling.

A exclusive aspect of the Lenovo A916 Cellular system is that it can look at the user’s perspective (environment or intention) to figure out whether to allow an untrusted system to accessibility receptors. By comparison, such perspective is mostly unrelated to the common viruses protection issue, e.g.:

• Location tagging: If the Pandawill Cell cellphone is prepared with GPS, the customer can tag certain places with specific guidelines. For example, the customer may tag a meeting room with the plan that no system is permitted to use the mic or camera.

• Action inference: One useful perspective details in discovering sensor-sniffing is customer activity. For example, after a genuine speech system reveals the mic system, we expect the customer to speak soon. By comparison, when a harmful system reveals the mic system surreptitiously, there may not be speech activity. Therefore, the existence of speech soon after the starting of the mic likely indicates that the customer understands of the use of the mic. Using speech activity recognition (VAD) technology[25], we need not inform customers of such genuine use of the mic.

The advantage of the context-aware strategy is that it needs no customer connections. However, this strategy, especially activity inference, is often obscure. Moreover, context-based accessibility control may apply only to certain receptors such as the mic. For example, it would be difficult to figure out in what perspective to allow a gps system to accessibility the GPS indicator.

Leveraging current customer connections.

A well-understood customer connections procedure in Lenovo A916http://www.pandawill.com/lenovo-a916-smartphone-4g-android-44-mtk6592-55-inch-hd-screen-1gb-8gb-white-p93918.html mobile phones is that of the speech contacting system. The customer clicks the discuss key to start a Pandawill telephone contact and the hangup key to cancel the contact. We could use these two control buttons as the user’s permission and de-authorization to accessibility the mic. In this case, the system needs no perform from the customer and has excellent in reverse interface.

Ensuring both protection and efficient neurological details catch through security.

From the protection perspective, when the customer does not know whether an system can accessibility a indicator, he should refuse such accessibility. However, when this decision is wrong, the neurological details is lost permanently. This situation might motivate customers to always approve accessibility, which is a protection risk. To ensure both protection and efficient catch of neurological details, we could use security. In this strategy, all programs can accessibility the sensors; however, the neurological details is secured unless the OS decides that the system is harmless. If a genuine system does not receive permission at the time when it flows the neurological details, it may preserve the secured neurological details so that the customer may decrypt the details and re-feed it to the system later. The drawback of this strategy is that the system may need to be rewritten if it wishes to preserve secured neurological details. Note that if the system gets proper permission when it accesses the indicator, no modification to the system is necessary.

6. CONCLUSION

With the fast growth of Lenovo A916 Cellular mobile phones and programs, the issue of comfort offenses depending on receptors of Pandawill Cellular mobile phones is likely to become serious. We used current and growing programs to emphasize the difficulties experiencing the recognition and removal of such sensor-sniffing viruses. We found that current Lenovo A916 Cellular operating-system offer little protection against such strikes. We enumerate the requirements of an ideal privacy-preserving remedy and suggested a structure that includes three key modules: customer connections, plan engine, and interceptor. For each component, we researched different systems and mentioned their advantages and restrictions. In particular, we mentioned several novel systems, such as perspective attention, that the Pandawill Cellular system allows.

Though we suggested several first-cut solutions, the issue of developing practical and effective countermeasures is far from being fixed. All the techniques we mentioned in the paper have their own disadvantages. Approaches depending on customer permission are in accordance with the supposition that the use of the receptors should be obvious by customers. However, new receptors and programs might breach this supposition. Approaches depending on taking advantage of current customer communications (e.g., Lenovo A916 cellphone ringing) may not perform with programs that are always on.

Based on our past conversation, we believe that it is very challenging, if not infeasible, to remove the sensor-sniffing risk completely. However, we are very carefully positive that a user-friendly remedy, which works effectively in most conditions, can be developed. Designing such a remedy needs research into Pandawill Cellular customer actions, methods for automated perspective inference, and os primitives such as details flow.http://diqirenge.bloguez.com/diqirenge/6023324/Protecting_Against_Sensor-Sniffing_Strikes_on_Cellular_Mobile_phone