There has been much research focused on securing wireless sensor network. One of the most difficult issues related to security in the wireless sensor network environment is how to minimize the impact of node-compromised problems.
In [7], K. Bıçakcı, C. Gamage, B. Crispo, and A. Tanenbaum proposed one-time sensors mechanism to mitigate node-capture attacks. Below is the description of their mechanism.
First, the base station preloads every sensor node with a unique ID value and a single cryptographic token. All sensor nodes are also preloaded with a sufficient amount of verification data to enable them to check the validity of tokens received. In every node there is also a memory space reserved to store the revocation list, which is initially empty.
Then the operation is performed as follows：
1. Based on its local routing information, when an one-time sensor senses the target event, such as a fire, it sends an alarm message to one node or multiple nodes.
Through this routing path, the alarm message can be sent to the base station. The alarm message is basically consists of the ID and cryptographic token of the sensing node.
2. When a node receives an alarm message, it would first check if it has already received a valid alarm message from the same node by comparing the ID value with the entries in its revocation list. If yes, it had received the other alarm message from the node which generated the alarm message. If not, it then ensures that the cryptographic token it received is indeed valid. Only if the cryptographic token is verified correctly, then the following two actions are taken. First, the alarm message is forwarded to the node(s) on the way to the base station. Second, the ID of the sender is added to the
revocation list for future reference.
3. The second step repeats itself with other nodes until the alarm message is received by the base station. The base station verifies that the alarm message is valid and has not been received before. Based on the threshold value and the number of messages it received before, the base station either decides to notify an alarm or waits for additional alarm messages.
From the above description, we know that when a sensor node detects an event and sends the alarm message out, the forwarding nodes can deliver this alarm message at most once. It means that each sensor can be used only one time. Therefore, the applicability of this method has great restriction. It only suits the applications where the sensors can be used only one-time, such as sensors detecting a fire sense, nuclear attacks, and chemical outbreak.
In [1], Z. Yanchao, L. Wei, L. Wenjing, and F. Yuguang developed a location-based threshold-endorsement scheme to thwart the bogus data that attackers use the captured node to inject. This mechanism is based on the Tate Paring technique on Elliptic Curve Cryptography to provide authentication, key establishment, endorsement and verification.
It can prevent malicious nodes from joining the wireless sensor network and diminish the node-compromised problem efficiently. Since sensors need to compute the Elliptic Curve Cryptography operations when they sign or verify messages, it is comparatively too power-consuming.
Z. Yanchao, L. Wei, L. Wenjing, and F. Yuguang in [1] thought that the Tate Paring technique can be workable in wireless sensor networks because K. Bıçakcı, C. Gamage, B.
Crispo, and A. Tanenbaum in [8] computed the Tate Pairing with the similar parameters as theirs. Also, these researchers in [8] quoted that the execution cost of the Tate paring
operation was only 62.04ms and 25.5mJ.
However, G. Bertoni, L. Chen, P. Fragneto, K. Harrison, and G. Pelosil in [8]
implemented the Tate Paring operation in 32-bit ST22 smartcard microprocessor at 33MHz.
The sensors presently do not have the chip of the Tate paring operation. If this scheme implements on current WSNs, it is unclear whether the low-microprocessor sensor nodes can sign messages together quickly, verify them fast, and send them to the base station immediately or not. If the base station cannot be notified immediately without any delay, the monitor system is, in fact, useless. Additionally, sensor nodes rely on batteries as their source of power. The power-consuming Elliptic Curve operation is easy to make sensors have no power, which, in turn, would lead sensor nodes become ineffective. For these reasons, we doubt if this scheme in [1] is suitable for use in the current WSN.
Some researchers in [2]-[6] discussed the feasibility of using the public-key cryptography architecture, such as RSA or Elliptic Curve Cryptography, in WSNs. The investigators in [5] thought that if the public-key cryptography architecture should be feasible in WNSs, then the sensor nodes can embed the chip with the operation of public-key cryptography. However, sensor nodes embed the chip of tamper resistant hardware better than the chip of these public-key operations. When attackers invade the sensor nodes with tamper resistant hardware, these sensor nodes can prevent their data from being obtained. Thus, we do not need to consider node-compromised attacks, and simple secure mechanisms can guarantee the security in networks.