Reducing Authentication Signaling Traffic in
Third-Generation Mobile Network
Yi-Bing Lin, Fellow, IEEE, and Yuan-Kai Chen
Abstract—In universal mobile telecommunication system
(UMTS), authentication functions are utilized to identify and authenticate a mobile station (MS) and validate the service request type to ensure that the user is authorized to use the particular network services. The authenticating parties are the authentica-tion center (AuC) in the home network and the MS. In UMTS, the serving general packet radio service support node (SGSN) accesses the AuC to obtain the authentication data, and delegates the AuC to perform mutual authentication with the MS. Since the cost for accessing AuC is expensive, the SGSN may obtain an array of authentication vectors (AVs) at a time so that the number of accesses can be reduced. On the other hand, if the size of the AV array is large, the AV array transmission from the AuC to the SGSN may be expensive. Thus, it is desirable to select an appropriate value to minimize the authentication network signaling cost. We propose an analytic model to investigate the impact of on the network signaling traffic, which is validated by simulation experiments. Then, we propose an automatic -selection mechanism that dynamically selects the size of the AV array to reduce the network signaling cost. Our study indicates that the automatic -selection mechanism effectively identifies appropriate size of the authentication vector array.
Index Terms—Authentication, authentication center (AuC),
gen-eral packet radio service (GPRS), home location register (HLR), mobility management, serving GPRS support node (SGSN), uni-versal mobile telecommunication system (UMTS).
NOMENCLATURE
ADR Authentication data request and response. AuC Authentication center.
AUTN Authentication token. AV Authentication vector. CDR Call detail record.
CK Cipher key.
GGSN Gateway GPRS support node. GPRS General packet radio service. HLR Home location register.
IK Integrity key.
IMSI International mobile subscriber identity.
MS Mobile station.
OMC Operation and maintenance center.
Manuscript received May 16, 2001; revised September 21, 2001; accepted October 15, 2001. The editor coordinating the review of this paper and ap-proving it for publication is Y. M. Fang. This work was supported in part by MOE Program of Excellence Research under Contract 89-E-FA04-4, TAHOE Network, Ericsson, InterVideo, FarEastone, National Science Council under Contract NSC 90-2213-E-009-156, and by the Lee and MTI Center for Net-working Research, NCTU.
The authors are with the Department of Computer Science and Information Engineering, National Chiao Tung University, Hsinchu, Taiwan, R.O.C. (e-mail: [email protected]).
Digital Object Identifier 10.1109/TWC.2003.811171
RA Routing area.
RAND Random number.
SGSN Serving GPRS support node.
UAR User authentication request and response. UMTS Universal mobile telecommunication system. USIM User service identity module.
UTRAN UMTs terrestrial radio access network. XRES Expected response.
I. INTRODUCTION
U
NIVERSAL mobile telecommunication system (UMTS) is a third-generation (3G) mobile service technology evolved from general packet radio service (GPRS), which sup-ports multimedia services to the mobile users [12], [15]–[17]. The UMTS architecture is illustrated in Fig. 1. In this archi-tecture, the packet data services of an mobile station (MS) are provided by the serving GPRS support node (SGSN) con-necting to the UMTS terrestrial radio access network (UTRAN) that covers the MS. For the discussion purpose, we refer to the area covered by an SGSN as the SGSN service area. The SGSN connects the MS to the external data network through the gateway GPRS support node (GGSN). Furthermore, the SGSN communicates with the home location register (HLR) and the authentication center (AuC) to receive subscriber data and authentication information of an MS. The HLR maintains the current location (SGSN number) of an MS. The AuC is used in the security data management for the authentication of subscribers. The AuC may be colocated with the HLR. More details of UMTS and GPRS can be found in [4] and [13].An SGSN service area is partitioned into several routing areas (RAs). When the MS moves from one RA to another, a location update is performed, which informs the SGSN of the MSs cur-rent location. Note that a crossing of two RAs within an SGSN area requires an intra-SGSN location update, while a crossing of two RAs of different SGSN areas requires an inter-SGSN lo-cation update. Details of lolo-cation update and mobility manage-ment modeling can be found in [5] and [6].
In UMTS, authentication function identifies and authenti-cates an MS, and validates the service request type to ensure that the user is authorized to use the particular network services. Specifically, authentication is performed for every location update (either inter-SGSN or intra-SGSN), call origination, and (possibly) call termination. UMTS authentication supports mutual authentication, i.e., authentication of the MS by the network and authentication of the network by the MS. The procedure also establishes a new UMTS cipher key (CK) and integrity key (IK) agreement between the SGSN and the MS. In
Fig. 1. Simplified UMTS architecture.
Fig. 2. UMTS authentication procedure.
UMTS authentication, the authenticating parties are HLR/AuC in the home network and the user service identity module (USIM) in the user’s MS. Two major authentication procedures are described in this paper.
Distribution of Authentication Vector: This procedure dis-tributes authentication vectors (AVs) from the HLR/AuC to the SGSN. An AV is temporary authentication data, which enables an SGSN to engage in UMTS authentication and key agreement with a particular user. We assume the following.
• The HLR/AuC trusts that the SGSN will handle authenti-cation information securely.
• The communication path between the SGSN and HLR/AuC are adequately secure.
• The user (MS/USIM) trusts the HLR/AuC.
Authentication and Key Establishment: GSM only provides one-way authentication. In UMTS, mutual authentication is achieved by sharing a secret key between the USIM and the HLR/AuC [3]. This procedure follows a challenge/response protocol identical to the GSM subscriber authentication and key establishment protocol [13] combined with a sequence-number-based one-pass protocol for network authen-tication derived from ISO/IEC 9798-4 [7].
Signaling flows of above two procedures are described in the following steps (see Fig. 2).
Step 1) When an MS moves into a new SGSN area, the SGSN does not have previously stored authenti-cation information (i.e., AVs). The SGSN invokes the distribution of authentication vector procedure by sending the Authentication Data Re-quest message to the HLR/AuC. This message includes the international mobile subscriber identity (IMSI) that uniquely identifies the MS.
Step 2) Upon receipt of a request from the SGSN, the IMSI is used to identify the HLR/AuC record of the MS, and the HLR/AuC sends an ordered array of AVs (generated based on the MS record) to the SGSN through the Authentication Data Response message. An AV consists of a random number RAND, an expected response XRES, a cipher key CK, an integrity key IK, and an authentication token AUTN. Each AV is good for one authentication and key agreement between the SGSN and the USIM.
The HLR/AuC may have precomputed the re-quired AVs and retrieve these AVs from the HLR database or may compute them on demand. The next two steps authenticate the user and establish a new pair of cipher and integrity keys between the SGSN and the USIM.
Step 3) When the SGSN initiates an authentication and key agreement, it selects the next unused authentication vector from the ordered AV array and sends the pa-rameters RAND and AUTN (from the selected au-thentication vector) to the USIM through theUser Authentication Requestmessage.
Step 4) The USIM checks whether AUTN can be accepted and, if so, produces a response RES which is sent back to the SGSN through the User Authen-tication Response message. The SGSN compares the received RES with XRES. If they match, then the authentication and key agreement exchange is successfully completed. Note that in this mutual authentication procedure, AUTN is used by the USIM to authenticate the network, and RES/XRES is used by the network to authenticate the USIM.
In this step, the USIM also computes CK and IK using the received AUTN. During data delivery, CK and IK are utilized to perform ciphering and integrity functions in the MS. On the other hand, the SGSN retrieves CK and IK from the AV and passes them to the UTRAN for data ciphering and integrity. Details of CK and IK generation at the MS side are given in [3, Secs. 6.5, 6.6].
Note that the message names in the above descriptions are based on [3]. In [1], the actual SS7 messages sent in Steps
Fig. 3. Timing diagram.
MAP_SEND_AUTHENTICATION_INFO_ack. In [2], the messages exchanged in Steps 3 and 4 areAuthentication RequestandAuthentication Response.
Third-generation partnership project (3GPP) TS 33.102 [3] describes a procedure to distribute authentication data from a previously visited SGSN to the newly visited SGSN. This pro-cedure assumes that the links between SGSNs are adequately secure. In reality, it may not be true (especially when SGSNs are owned by different service providers or even different coun-tries). Thus, we assume that when the MS moves from the old SGSN area to a new SGSN area, the authentication data stored in the old SGSN are not sent to the new SGSN. Instead, the new SGSN obtains a new AV array from the HLR/AuC through anAuthentication Request and Response(ADR) message pair exchange (that is, Steps 1 and 2 in Fig. 2).
Consider the timing diagram in Fig. 3. Suppose that an MS enters a new SGSN area at time . The MS sends a registration message to the SGSN. Since the SGSN does not have authentication information, the distribution of authenti-cation vector procedure is executed through an ADR, where the number of AVs obtained from the HLR/AuC is . After the SGSN has obtained the AV array from the HLR/AuC, mutual authentication is performed between the SGSN and the MS/USIM through a User Authentication Request and Response(UAR) message pair exchange (that is, Steps 3 and 4 in Fig. 2) by using the first AV. After , the second authentication event (a call request or an inter-SGSN RA update) occurs at time . The MS/USIM initiates the second UAR, and the SGSN uses the second AV in the array for mutual authentication. At time , the last AV in the array is used for the UAR of the th authentication event. After , the next authentication event occurs at . The SGSN realizes that no AV is available, which issues the second ADR to obtain the next AV array from the HLR/AuC and then performs a UAR. For the next incoming authentication events, ADRs and UARs are executed accordingly as described above. At time , the MS leaves the SGSN area. The last authentication event before
occurs at (where ), which utilizes the
th AV in the AV array. Thus, during the period , UARs and ADRs are performed, and at time , AVs are left unused in the SGSN.
We note that the ADR operation is expensive (especially when SGSN and HLR/AuC are located at different countries). Therefore, one may increase the AV array size to reduce the
number of ADRs performed when an MS is in an SGSN area. On the other hand, with a large , the AVs may occupy too much network bandwidth for every transmission from the AuC to the SGSN. Thus, it is desirable to select an appropriate value to minimize the authentication network signaling cost. In the next section, we propose an analytic model to investigate the impact of on the network signaling traffic. Based on the analysis, we then propose a mechanism to select appropriate size of the authentication vector array. The notation used in this paper is listed in the Nomenclature.
II. ANALYTICMODELINGWITHFIXED
Let be the total number of ADRs performed when the MS resides in an SGSN area. For each ADR, the number of AVs obtained from the HLR/AuC is . Suppose that the ag-gregate incoming/outgoing call and registration arrivals form a Poisson process with rate . As we mentioned earlier, for every incoming/outgoing call and registration, a UAR is performed. For a specific period , let be the probability that there are ADRs to the HLR/AuC. Note that ADRs are
per-formed if there are ( ) UARs in the
period . According to the probability function of the Poisson distribution, we have
(1)
Let be the period that an MS resides in an SGSN service area. That is, in Fig. 3. Suppose that has a general distribution with the density function , the mean , and
the Laplace transform . Let
be the probability that there are ADRs during the MSs resi-dence in the SGSN area. Then
We derive and based on three SGSN residence time distributions as follows.
Gamma Distribution: If is a gamma density function with mean and variance , then
and
(5) We are particularly interested in the gamma distribution. It has been shown that the distribution of any positive random variable can be approximated by a mixture of gamma distributions (see [8, Lemma 3.9]). One may also measure the SGSN residence times of an MS in a real PCS network and the measured data can be approximated by a gamma distribution as the input to our analytic model. It suffices to use the gamma distributions with different shape and scale parameters to represent different SGSN residence time distributions [14].
From (5), (3) is rewritten as
(6)
Hyper-Erlang Distribution: Another distribution that has very general approximation capability to the probability distri-bution of any nonnegative random variable is the hyper-Erlang distribution [6], [8] with the mean
where is a positive integer, and
The Laplace transform of the hyper-Erlang distribution is (7)
that is either Coxian or SOHYP distributed or more generally phase-type distributed [9], [10].
Exponential Distribution: Let in (6) or let in (8). In this case, is exponentially distributed, and (3) is rewritten as
Also, (4) is rewritten as
(9)
The exponential distribution may not capture the details of the real SGSN residence times for a mobile network. However, this distribution is good for mean value analysis, which does capture the trend for the performance of a system [11].
Let be the total message transmission cost for ADRs when an MS resides in an SGSN area. Then
(10) where represents the cost for an SS7 message overhead nor-malized by the cost of an AV transmission and the processing time for generating the AV through the authentication proce-dure. In the right-hand side of (10), this overhead is considered for theRequestandResponsemessage pair exchanged in an ADR. From (9) and (10), the total ADR transmission cost for exponential SGSN residence times is expressed as
(11)
III. PERFORMANCE OFFIXED- MECHANISM
This sections shows how the expected number and cost of ADR are affected by the AV array size . We say that the AV array size is selected by the fixed- mechanism if the selected value is fixed throughout an MS’s lifetime. Our ana-lytic model in the previous section can be used to compute and for the fixed- mechanism. The analytic results are compared with the simulation experiments as shown in Figs. 4 and 6. In both figures, the dashed curves represent the analytic results, and the symbols , , , and represent the simula-tion results. These figures indicate that the analytic results are consistent with the simulation experiments. Fig. 4 plots against with various UAR (authentication event) arrival rates
Fig. 4. Effect of the UAR arrival rate (exponential SGSN residence times with mean1=).
Fig. 5. P (n; K) probability distribution (exponential residence times; =
10).
distributed with mean . It is obvious that is a de-creasing function of . The figure indicates that for
and , is only insignificantly reduced by in-creasing . Fig. 5 plots the probability distribution, where . The figure shows that the shapes of
distribution are similar for , which are very different from the shape when . This observation is consistent with our statement for Fig. 4; i.e., the values are roughly the same for large values, and increasing does not improve
the performance.
Fig. 6 shows the effect of the variance for the gamma SGSN residence time distribution. The mean SGSN residence time is , and the UAR arrival rate is . The figure shows that as increases, increases. This phenomenon is explained as follows. As increases, more short SGSN residence times and more long SGSN residence times are observed. For short residence times, it is likely that in Fig. 3 and, thus, is expected. Fig. 7 indicates that signif-icantly increases as increases (especially when ).
Fig. 6. Effect of the variancev for gamma SGSN residence times ( = 10).
Fig. 7. Effect of the variancev for gamma SGSN residence time distribution ( = 20).
In this case, it implies that after the SGSN has obtained the AV array, only one AV is used for the first UAR. The AVs are wasted. Similarly, as increases, more long residence times are observed, and a larger ADR number is expected [see
in Fig. 7]. Thus, is an increasing function of . The above discussion leads to the conclusion that for MSs with irregular
Fig. 8. Effect of the SGSN residence time distribution.
moving patterns, more network signaling traffic for authentica-tion will be observed. As a final remark for Fig. 7, more resi-dence times with lengths close to its mean value are observed for smaller variance [see in Fig. 7].
Fig. 8 plots when the SGSN residence time distribu-tions are gamma and Erlang. In our example, the
hyper-Erlang distribution has parameters , ,
, , and . From the above
param-eters, we have
(12)
In Fig. 8, we set , and . The ,
and values for the case of gamma distribution are assigned based on (12). Note that if , then as decreases, in-creases and decreases. Thus, the curves are decreasing (because decreases as increases), which are different from what we observed in Fig. 6, where is fixed. Fig. 8 indicates that curves are similar for gamma and hyper-Erlang dis-tributions. In the remainder of this paper, we only consider the gamma SGSN residence time distribution for our performance study. The results for hyper-Erlang residence times are similar and will not be presented.
By assuming in (10), Fig. 9 plots against with various UAR arrival rates . The SGSN residence times are ex-ponentially distributed in Fig. 9(a), and are gamma distributions with various variance values in Fig. 9(b). Since is a de-creasing function of (see Figs. 4 and 6), the product
[and thus ] results in concave curves shown in Fig. 9. That is, as increases, decreases and then increases. Let be the optimal value that minimizes the cost . For exponential
IV. AUTOMATIC -SELECTIONMECHANISM
In 3GPP TS 29.002 [1], is recommended for fixed-mechanism. However, our study in the previous section indi-cates that the optimal value is affected by the traffic of UARs. This section proposes a per-user automatic -selection mechanism that dynamically selects the value based on the SGSN residence time distribution and the UAR traffic pattern of an MS. The automatic -selection mechanism can be implemented in MS or in AuC. If the mechanism is exercised at the MS, then every time the MS moves into a new SGSN service area, it forwards the new value to the AuC through the registration procedure. If the mechanism is exercised by the AuC, then the number of the UARs must be provided by the old SGSN when the MS moves to the new SGSN or when the MS detaches from the network. The old SGSN can either supplies the number of the UARs (the number of calls and intra-SGSN registrations when the MS resides in the old SGSN) to the HLR/AuC when theCancel Locationmessage pair is ex-changed between the old SGSN and the HLR (see [4, Sec. 6.9] or [13, Sec. 2, Ch. 9]), or the SGSN can generate call detail records (CDRs) for calls and intra-SGSN location updates (see [13, Sec. 6, Ch. 18]). The CDRs are sent to either a billing gateway or an operation and maintenance center (OMC), and the OMC forwards the related information to the HLR/AuC.
Let be the value selected for the th iteration (i.e., when the MS resides in the th SGSN area). The automatic
-selection mechanism is described as follows.
Initialization: When the user subscribes to the UMTS ser-vice, an initial value is assigned (e.g., as 3GPP suggested). Then, every time the MS enters an SGSN area, the following two steps are executed.
Measurement Step: During the th SGSN residence period, the number of UARs are counted.
Decision Step: When the MS leaves the th SGSN area, we determine if the value should be adjusted to
(decrement by one), (no change),
(increment by one). We use (10) as the heuristic to compute three costs as follows:
for (13)
Based on (13), is selected as
(a) (b)
Fig. 9. Cost of ADRs (the mean of the SGSN residence times is1=). (a) Exponential residence times. (b) Gamma residence times ( = 10).
Fig. 10. Performance of the K-selection mechanism: Exponential SGSN residence times with mean1=.
From the above discussion, it seems reasonable that when the MS enters the st SGSN area, the AV array size for ADRs is assigned the value computed in (14). However, from our experiments (to be elaborated later), this is not the best choice. Instead, the best performance is achieved when a smaller value is selected. Define three AV array size assignments for the st iteration as follows.
: AV array size is .
: AV array size is .
: AV array size is .
As we will see later, will yield the best performance in most cases. Also note that choosing any AV array sizes larger than always yields performance worse than that for and these scenarios will not be presented in this paper.
Figs. 10–12 show the performance of , , and , and compares them with the fixed- mechanism. We assume that in (13). Note that the actual value depends on the implementation of the AV data structure in a mobile network. However, the conclusions presented in this paper holds for any
Fig. 11. Performance of the K-selection mechanism: Gamma SGSN residence times ( = 10).
Fig. 12. Performance of the K-selection mechanism: Mixed UAR arrival patterns.
values. Let , , and be the ADR transmission cost for , , and , respectively. As defined before, let be the cost of fixed- mechanism with the array size . Three
where is the optimal value that minimizes the cost in the fixed- mechanism.
Gamma SGSN Residence Time Distribution: Fig. 11 plots the cost curves, where the UAR arrival rate is , and the SGSN residence times have gamma distribution with variances , , and , respectively. Inequality (15) also holds for most cases in this scenario. Although the cost for is slightly greater than that for when is large, the difference is insignificant and can be ignored. However, becomes less accurate in terms of capturing the value. The cost of is very close to as in the case for exponential SGSN residence times.
Mixed UAR Arrival Patterns: Fig. 12 plots the cost curves where the UAR arrivals have mixed patterns. In this figure, the solid curve represents the case where the SGSN residence times have an exponential distribution with mean . The dashed curve represents the case where the residence time distribution
is gamma with mean and variance . In both
curves, the UAR arrivals are Poisson with rate for the first 10 000 UARs, then the rate changes to for the second 10 000 UARs, for the third 10 000 UARs, and
for the last 10 000 UARs.
The figure indicates that with the mixed UAR traffic patterns, and may outperform the fixed- mechanism for all values (see the solid curve).
The above analysis indicates that the performance of the auto-matic -selection mechanism (particularly ) is close to (and may be better than) the performance of the fixed- mechanism with the optimal value.
V. CONCLUSION
In the mobile network, authentication is exercised for every location update and every call event. In UMTS, mutual authen-tication is performed between the MS and the SGSN. In order to carry out authentication, the SGSN obtains authentication data from the HLR/AuC. Since the cost for accessing AuC is expen-sive, the SGSN may obtain an array of AVs at a time so that the number of accesses can be reduced. On the other hand, if the size of the AV array is large, every AV array transmission from the AuC to the SGSN may be expensive. Thus, it is desir-able to select an appropriate value to minimize the authenti-cation network signaling cost. We proposed an analytic model to investigate the fixed- mechanism (the selected value is fixed throughout an MS’s lifetime). The analytic results were validated against the simulation experiments. We observed the following results.
(UAR) rate increases.
Study on the fixed- mechanism suggests that should be ad-justed based on the authentication traffic so that the cost is reduced. We proposed a per-user automatic -selection mech-anism that dynamically selects the value based on the SGSN residence time distribution and the UAR traffic pattern of an MS. Our study showed that the automatic -selection mech-anism effectively identifies appropriate value to reduce the network signaling cost. Specifically, our study indicated that the performance of the automatic -selection mechanism (partic-ularly ) is close to (and may be better than) the performance of the fixed- mechanism with the optimal value.
ACKNOWLEDGMENT
The authors would like to thank the three anonymous reviewers. Their comments have significantly improved the quality of this paper.
REFERENCES
[1] 3GPP, 3rd Generation Partnership Project; Technical Specification Core Network; Mobile Application Part (MAP) Specification (Release 1999), 2000, Technical Specification 3G TS 29.002 V3.7.0 (2000-12). [2] 3GPP, 3rd Generation Partnership Project; Technical Specification
Group Core Network; Mobile Radio Interface Layer 3 Specification; Core Network Protocols—Stage 3 for Release 1999, 2000, 3G TS 24.008 version 3.6.0 (2000-12).
[3] 3GPP, 3rd Generation Partnership Project; Technical Specification Group Services and Systems Aspects; 3G Security; Security Architec-ture, 2000, Technical Specification 3G TS 33.102 V3.7.0 (2000-12). [4] 3GPP, 3rd Generation Partnership Project; Technical Specification
Group Services and Systems Aspects; General Packet Radio Service (GPRS); Service Descripton; Stage 2, 2000, Technical Specification 3G TS 23.060 version 3.6.0 (2001-01).
[5] I. F. Akyildiz, J. McNair, J. S. M. Ho, H. Uzunalioglu, and W. Wang, “Mobility management in next generation wireless systems,” Proc. IEEE, vol. 87, pp. 1347–1384, Aug. 1999.
[6] Y. Fang and I. Chlamtac, “Teletraffic analysis and mobility modeling for PCS networks,” IEEE Trans. Commun., vol. 47, pp. 1062–1072, July 1999.
[7] ISO/IEC, “Information technology-security techniques—Entity authen-tication—Part 4: Mechanisms using a cryptographic check function,” ISO/IEC, Tech. Rep. ISO/IEC 9798-4, 1999.
[8] F. P. Kelly, Reversibility and Stochastic Networks. New York: Wiley, 1979.
[9] L. Kleinrock, Queueing Systems: Volume I—Theory. New York: Wiley, 1976.
[10] G. Latouche and V. Ramaswami, Introduction to Matrix Analytic Methods in Stochastic Modeling. Philadelphia, PA: SIAM, 1999. [11] E. D. Lazowska, J. Zahorjan, G. S. Graham, and K. C. Sevcik,
Quantita-tive System Performance. Englewood Cliffs, NJ: Prentice-Hall, 1984. [12] B. Li et al., “QoS-enabled voice support in the next-generation Internet: Issues, existing approaches and challenges,” IEEE Commun. Mag., vol. 38, no. 4, pp. 54–61, Apr. 2000.
[13] Y.-B. Lin and I. Chlamtac, Wireless and Mobile Network Architec-tures. New York: Wiley, 2001.
[14] Y.-B. Lin, W. R. Lai, and R. J. Chen, “Performance analysis for dual band PCS networks,” IEEE Trans. Comput., vol. 49, pp. 148–159, Feb. 2000.
[15] UMTS Forum. (2000) Enabling UMTS/third generation services and applications. Tech. Rep. 11, UMTS. [Online]. Available: www.umts-forum.org
[16] UMTS Forum. (2000) Shaping the mobile multimedia future—An ex-tended vision from the UMTS forum. Tech. Rep. 10, UMTS. [Online]. Available: www.umts-forum.org
[17] UMTS Forum. (2000) The UMTS third generation market—Structuring the service revenues opportunities. Tech. Rep. 9, UMTS. [Online]. Available: www.umts-forum.org
[18] E. J. Watson, Laplace Transforms and Applications. Cambridge, MA: Birkhauserk, 1981.
Yi-Bing Lin (M’96–SM’96–F’03) received the
B.S.E.E. degree from National Cheng Kung Uni-versity, Taiwan, in 1983 and the Ph.D. degree in computer science from the University of Wash-ington, Seattle, in 1990.
From 1990 to 1995, he was with the Applied Research Area, Bell Communications Research (Bellcore), Morristown, NJ. In 1995, he was appointed Professor of computer science and infor-mation engineering (CSIE), National Chiao Tung University, Hsinchu, Taiwan, R.O.C. (NCTU). In 1996, he was appointed Deputy Director of Microelectronics and Information Systems Research Center, NCTU. From 1997 to 1999, he was elected as Chairman of CSIE, NCTU, Hsinchu, Taiwan, R.O.C. He is an Editor for Computer Networks, an Area Editor of ACM Mobile Computing and Com-munication Review, a columnist of ACM Simulation Digest, an Editor of the International Journal of Communications Systems, an Editor of ACM/Baltzer Wireless Networks, an Editor of Computer Simulation Modeling and Analysis, an Editor of the Journal of Information Science and Engineering, and Guest Editor for the ACM/Baltzer MONET Special Issue on Personal Communica-tions. He is the coauthor of Wireless and Mobile Network Architecture (New York: Wiley, 2001). His current research interests include design and analysis of personal communications services network, mobile computing, distributed simulation, and performance modeling.
Dr. Lin is an Associate Editor of the IEEE NETWORKING, an Editor of the IEEE TRANSACTIONS ONWIRELESSCOMMUNICATIONS, an Editor of the IEEE Personal Communications Magazine, a Guest Editor for IEEE TRANSACTIONS ONCOMPUTERS: Special Issue on Mobile Computing, and a Guest Editor for IEEE Communications Magazine: Special Issue on Active, Programmable, and Mobile Code Networking. He is Program Chair for the 8th Workshop on tributed and Parallel Simulation, General Chair for the 9th Workshop on Dis-tributed and Parallel Simulation, and Program Chair for the 2nd International Mobile Computing Conference. He received the 1998 and 2000 Outstanding Research Awards from National Science Council, R.O.C., and the 1998 Out-standing Youth Electrical Engineer Award from CIEE, R.O.C. He is an Adjunct Research Fellow of Academia Sinica.
Yuan-Kai Chen received the B.S.C.S.I.E. and
M.S.C.S.I.E. degrees from National Chiao Tung University, Hsinchu, Taiwan, R.O.C., in 1989 and 1991, respectively. He is currently working toward the Ph.D. degree in the Department of Computer Science and Information Engineering, National Chiao Tung University.
In 1991, he joined the Telecommunication Laboratories, Chunghwa Telecom Company, Ltd., Taiwan, R.O.C. His current research interests include design and analysis of personal communications services network, mobile computing, and UMTS.
