Data outsourcing is economically attractive; however, data security, integrity, availability and many other concerns are elements or factors limiting the adoption of data outsourcing by potential users [1] . Many of these factors such as the data secrecy and availability were addressed in the field literature for long time. In addition, the emerging era of cloud computing imposes new challenges such as data monitoring [2] . Data monitoring could be based on real-time “intrusion detection” or on logging and offline auditing [3] . Many aspects of data outsourcing concerns in the cloud were addressed by researchers such as secrecy [4] [5] , integrity [6] [7] , availability [8] [9] .

Data storage and manipulation outsourcing in clouds presents one of the biggest clients’ concerns, especially for confidential data, concerns caused by the lack of tra- nsparency and limited user control; new attack channels for malicious users can exist due to the clouds multitenancy in conjunction with Virtual Machine Monitor (VMM) vulnerabilities [9] . Data monitoring techniques normally end up with logs that record the activities performed on data over time, such logs are the source for information that enable the process of verification of different aspects of auditing. A data-centric approach for control and auditing that applies across application and service boundaries seems appropriate for the demonstration of compliance with data regulations, and to understand how information is manipulated or generated [10] .

Auditing is the systematic evaluation to measure the adherence of a system―such as a cloud service―to the established criteria and pre-specified set of policies imposed by the client on his/her data; auditor is the party that can conduct independent assessments of cloud services. The audit processes have to be highly regarded as it is the shor- test way to gain users’ trust and satisfaction. Standardization entities such as United States National Institute of Standards and Technology (NIST) have identified the protection of audit data integrity as one of the major concerns for the cloud technology [3] .

Maintaining audit data integrity and security is challenging; for example, the CSP may try to hide any trails leading to uncover their responsibility about actions led to data misuse, losses, or security breaches [15] . It is very hard for the ordinary users―in most of the cases―to identify the reasons for such incidents under such circumstances [10] [11] . Therefore, a need for a frequent audit checking is needed to keep an eye on data, and to track the data movement to ensure only allowed flows exist [9] , and ensure that the CSPs meet the requirements of the policy being enforced on data [10] . There are no existing internal mechanisms for cloud services users to enforce policy or reliably demonstrate compliance of policies, law, and regulations [12] , furthermore, regarding the possibility of monitoring or audit, except for IBM and Microsoft Azure, service providers do not allow customers to hire a third party to perform such activities [13] .

Being able to identify the cause of a security breach and the path that has been followed by an attacker is much more difficult in cloud environment than other environments. In addition, a cloud provider may subcontract a third party for some resources or a hardware supplier of poor-quality storage devices that result in loss of data. Traditional investigation methods based on digital forensics cannot be extended to a cloud, since resources are shared among a large user population and the traces of events related to a security incident may be wiped out due to the high rate of write operations on any storage media.

The importance of contractual obligations between tenants and cloud providers related to outsourcing data storage and manipulation were addressed in the literature in the form of guiding notes and recommendations [13] - [17] . In [18] authors stated that although the CSPs provide security services to the clients’ data, there is a possibility of data leakage in cloud environment, users must make sure that the data moved into the cloud has not been leaked or modified by some unauthorized users, and to verify the security measures offered by the CSP.

The contractual terms between users and the CSPs make the user trust much worse; it usually places all responsibilities for data security on user’s side [3] . Maintaining copies of the data outside the cloud is often unfeasible due to the sheer volume of data. If the only copy of the data is stored on the cloud, sensitive data is permanently lost when cloud data replication fails and is followed by a storage media failure. Because some of the data often includes proprietary or sensitive data, access to such information by third parties could have severe consequences.

Close attention should be paid to both; the security of storage servers, and to data in transit. Data usually is kept in storage for extended periods of time and so the concerns of unauthorized access to confidential information and data theft are more higher when data is in storage than while it is being processed, data exposure to threats during processing is for relatively short periods of time.

The user control over the life cycle of data is also of an extreme importance. How can a user determine whether data that should have been deleted is actually deleted? Even if it was deleted, there is no guarantee that the media was wiped and the next media user is not able to recover any confidential data. The problem is magnified when the CSPs rely on seamless backups to prevent accidental data loss since it is done without users’ knowledge or approval. Such behavior may lead to data loss, accidental deletion, or data becomes accessible to an attacker.

In real world, two auditing schemes exist, private auditability and public auditability. In private auditability only the client or the data owner is allowed to check the different aspects of the stored data (i.e. secrecy, integrity, etc.), but it is not applicable for all clients and increases verification overhead of the user. Public auditability allows some entity privileged by the data owner to perform data verification check. This entity is called a Third Party Auditor (TPA). TPA has the expertise, capabilities, knowledge and professional skills that client normally does not have. TPAdoes frequent checks against data using different aspects and it reduces the client overhead; the client is no longer required to verify the integrity of the data at the server on his/her own. For data owners to have sufficient proficiency of trust; TPA may have only an encrypted format of the data, and are trusted by the data owner for auditing of the data and to ensure there is no data leakage and verify the compliance of the CSP with the contract.

Figure 1 [1] [19] shows public auditing scheme, in which, cloud users (clients) use the cloud storage provided by the CSP to outsource storage of their data, a TPA entity trusted by the client is entitled to keep checking on user’s data periodically to verify data integrity. In the case of any unexpected results, both the client and the CSP are notified, a flow of security messages are exchanged between the three entities. In Figure 2 [11] [20] the public auditing scheme is extended to support public verification and dynamic data auditing, it also supports batch auditing in order to improve efficiency. To overcome data privacy issues, random mask technology is used to avoid TPA getting the client's data information knowledge on every verification process.

In [21] , authors proposed a solution where the cloud provider generates audit logs, by which the cloud tenant is able to monitor the enforcement of data location-related regulation. Audit logs confidentiality and integrity should be ensured together with trusted time stamping. Cloud customers should be able to access the logs of every action including VMs’ deployment; migrations; and shutdowns. The logs can also be used for the verification of the Cloud’s utilization.

In [22] , a software that is able to estimate the geolocation of itself, named VLOC, is introduced to verify the physical location of a VM on which the customer applications and data are stored. VLOC notifies users if the location is unauthorized. To get its

location, the software uses arbitrary web-servers as external landmarks, and employs network latency measurement for distance estimation. In the case of latency fluctuation, VLOC employs a machine learning technique in order to adapt.

Authors In [23] introduced a data driven usage control and provenance tracking approach, acts as a light-weight, transparent proxy between storage users and providers to enforce compliance constraints in cloud storage federation scenarios. The compliance constraints in their work were based on the enforcement of regulations, standards, and laws in cloud storage. Constraints were categorized into spatial, temporal, and qualitative restrictions on distribution and replication of data. The proxy intercepts all attempts to retrieve, store, or delete data from a cloud storage provider and forwards the corresponding events to the a policy decision point (PDP), which, based on previously deployed policies, decides on their execution and modification and finally sends its decision back to the proxy, the proxy then, based on the PDP’s decision, either executes, modifies, or drops the event. The problems with such design are the high overhead on the cloud infrastructure, the possible single point of failure, the required implementation of the proxy at the CSP side, and it does not address data tracking once it leaves the environment of the federation service. In addition, no audit logs are generated.

In [10] [12] [24] , authors introduced Information Flow Audit, as an approach for tracking information flows within cloud infrastructure. This builds upon CamFlow (Cambridge Flow Control Architecture), model for data-centric security in PaaS clouds. CamFlow enforces Information Flow Control (IFC) policy both intramachine at the kernel level, and inter-machine on message exchange to provide datacentric audit logs. Combining a continuously enforced data-centric security mechanism with meaningful audit enables tenants and providers to both meet and demonstrate compliance with their data management obligations. They outlined how IFC can be enforced in the cloud and discussed how audit data can be collected as an intrinsic part of IFC.

Up to our best knowledge, only little work was considering the users’ concerns other than data security, integrity, and availability. Recent surveys on cloud computing security and auditing [25] - [28] didn’t cover the concerns we talk about here, namely; spatial, temporal, and legal concerns. Although a little work was dedicated to enforce compliance of CSPs with data location regulations using auditing, the demonstration of the compliance of cloud providers to their contractual obligations is not yet convincing and more research is needed [29] . We are proposing an extension to the auditing models to include such concerns in order to increase the trust of data owners in the cloud environment.

The rest of this paper is organized as follows: Section 2 covers data owners’ concerns limiting their adoption of clouds in their businesses, in the form of a questionnaire and discussion of results, the proposed extension to auditing models is presented in section 3, and finally; the conclusion along with future work direction is drawn in section 4.

The cloud computing technology advantages are obvious; however, major challenges do exist such as the availability of services, data confidentiality, and auditability. Clouds in its different deployment and delivery models in Jordan are widely available along with good resources and infrastructure to support this technology, nevertheless, limited adoption among data owners. This limited adoption observation is common to both public and private sectors, but it is more obvious in the public sector.

In order to investigate data owners’ concerns which are limiting their adoption of clouds in their businesses, we have conducted a questionnaire with thirteen questions (appendix A). Questions were selected to cover the area of data storage and manipulation outsourcing since it is, in our opinion, the major concerns producing and affecting clouds’ area and adoption, such opinion were developed during multi discussions with groups of data owners in different sectors in Jordan. We intended to understand the reasons behind the major concerns limiting the shift towards the clouds, and how it can be tackled. The questionnaire also included questions for the data owners to express their additional important aspects related to the cloud environments.

The questioned sample was carefully chosen to represent: public sector, private sector, and the academic community. Total number of participants who answered the questionnaire is 23, all of them are from the information technology field, working with information systems of very large sizes, and have good background in cloud computing technologies. Persons from the academic field are researchers in the cloud computing technologies. It is worth saying that it wasn’t an easy task to find a big number of typical participants to answer the questionnaire in order to get useful feedback because; as we indicated above; we carefully selected participants with the above mentioned knowledge and hence the small number; 23.

As shown in Appendix A, the questionnaire is composed of thirteen questions: eleven “Yes/No” questions and two fill in the blank questions. The results of eleven “Yes/No” questions of the questionnaire are summarized in Table 1, while Table 2 summarizes the two “fill in the blank” questions.

Table 1 represents the percentage results of the participants’ answers (see Appendix B) with Yes or No to a pre-specified list of concerns if they were to move their data to the cloud. The concerns list included legal, spatial, and temporal aspects of data outsourcing. Notice that Table 1 shows participants’ concerns listed in descending order (highest Yes percentage to lowest Yes percentage). Table 2 represents the recurrence of additional aspects or concerns as listed by the sample, which in their opinion may generate major concerns for a data owner when outsourcing data storage and manipulation (Q8 and Q9).

Results of the questionnaire showed that 100% of the sample believes that data outsourcing will generate major concerns (Q1); most of the concerns are directly related to data security, availability, and location. 91% of the sample believes that specifying a judiciary system and analyzing how information has flowed across the system are two major concerns (Q7 and Q11). Users’ ability to control their data, and the lack of CSPs’ transparency, were among the major concerns of data owners. About 83% of the sample believes that cloud multitenancy forms a new attach channel, it is important to