Dynamic SLAs for Clouds

Abstract

In the Cloud domain, to guarantee adaptation to the needs of users and providers, Service-Level-Agreements (SLAs) would benefit from mechanisms to capture the dynamism of services. The existing SLA languages attempt to address this challenge by focusing on renegotiation of the agreement terms, which is a heavy-weight process, not really suitable for dealing with cloud dynamism. In this paper, we propose an extension of SLAC, a SLA language for clouds that we have recently defined, with a mechanism that enable dynamic modifications of the service agreement. We formally describe this extension, implement it in the SLAC framework and analyse the impacts of dynamic SLAs in some applications. The advantages of dynamic SLAs are demonstrated by comparing their effect with that of static SLA and of the “renegotiation” approach

A Systematic Literature Study on Definition and Modeling of Service-Level Agreements for Cloud Services in IoT

“…Content may change prior to final publication. [29] 2017 -----Uriarte et al [30] 2014 -Uriarte et al [31] 2016 ----García et al [32] 2015 Schill et al [33] 2017 --Tata et al [14] 2016 --Papaemmanouil et al [34] 2014 -----…”

“…For instance, Serrano et al [28] present the Cloud Service Level Agreement (CSLA) language, while Hasan et al [29] extend this language to support Green SLAs. Similarly, Uriarte et al present the SLAC langauge for SLAs in [30] and its extension in [31]. Furthermore, the WSLA, S3LACC, RSLA and XClang languages for modelling SLAs are presented by García et al [32], Schill et al [33], Tata et al [14] and Papaemmanouil et al [34] respectively.…”

“…The analysis of the extracted data reveals that the degree of formalization in the SLA languages differs significantly between the contributions of the relevant papers as shown in Table 3. The degree of formalization in these works ranges from semi-structured documents written in natural language [28], [32], [34] to formal specification [14], [29]- [31], [33]. It is interesting to note that 62.5% of the total relevant publications on languages for modeling SLAs present formal contributions, whereas the remaining 37.5% of the publications provide semi-structured contributions.…”

“…We also note that not all contributions in the relevant publications on languages for SLA modeling are supported by tools as shown in Table 3. There are only three contributions [28], [30], [31] [23] 2018 -Rady et al [24] 2013 -----Labidi et al [25] 2017 ---Rady et al [26] 2014 -----Alhamad et al [13] 2010 -----Di Modica et al [27] 2012 -----…”

“…Finally, we note that apart from the two contributions [28], [31], none of the other contributions on languages for SLA modeling are fully mature according to the Redwine-Riddle model [16]. The contributions in [14], [29], [30], [32] are somewhat mature, while the contributions in [33], [34] are not mature as depicted in Table 3.…”