Cloud computing migrations are increasing rapidly. The main influencing factor being IT management costs. IoT-based enterprises that started their cloud journey by setting up small private clouds within their enterprise have often found that as the applications and services they use broaden. Then the shift towards incorporating public clouds becomes inevitable. The current problem that many of these firms are encountering is the difficulty of managing multiple clouds that reside within different vendors running on different platforms, computational requirements, and vendor SLAs. Lack of support for a single standard for an overall multi-cloud hybrid model exposes the hybrid IT-management to further threats. This makes it difficult for an adopting enterprise to manage and maintain its cloud-based systems during peak performance hours, which often leads to system downtime. This chapter discusses various SLA issues specific to a hybrid multi-cloud environment and suggests possible solutions to help adopting firms in their management.
TopHybrid Cloud Computing
Hybrid cloud computing is a widely accepted and implemented cloud model in both business and education sectors. Generally, enterprises use the hybrid model when they do not wish to completely outsource their entire IT infrastructure or processing to a remote location as this can increases the level of risk associated with data privacy, governance and control. Hybrid cloud offers various benefits to the Internet of Things (IoT) industry by overcoming the real-time data processing, storage, capacity, availability and storage. It is estimated that the number of IoT connected devices would increase to 50.1 billion by 2020, which may not allow the existing IT based architecture to function to its full capacity (Dhirani, Newe, & Nizamani, 2018a). IoT based tenants resisting cloud migrations may limit their applications reachability, availability and usability.
Cloud Classification
Cloud Computing can be categorized based on the architecture of models (Public cloud, Private cloud, Hybrid Cloud and Community Cloud), services (Infrastructure as a Service, Platform as a Service, Software as a Service, etc.) and applications provided by the cloud vendor on the cloud management portfolio. These models can also be categorized further such as: Lone Hybrid Cloud and Multi-Cloud Hybrid model.
Lone Hybrid Cloud
This model demonstrates a company outsourcing a part of its application and data processing with a single vendor at peak processing times. The relationship between the user and vendors can be formed in the following possible ways:
- •
1:1 (single tenant: single vendor)
- •
M:1 (multiple tenants: single vendor)
- •
M:1:M (multiple tenants: single vendor: multiple third-party subcontractors)
Multi-Cloud Hybrid Model
This model suits companies who develop multiple small private clouds and enter the hybrid environment with multiple vendors based on the vendors expertise and services. The relationship between the user and vendors can be formed in the following possible ways:
- •
1:M:M (single tenant: multiple private clouds: multiple public cloud vendors)
- •
M:M:M (multiple tenants, multiple private clouds: multiple public cloud vendors)
- •
M:M:M:M (multiple tenants, multiple private clouds: multiple public cloud vendors: multiple third-party subcontractors)
Cloud Sla Issues
Some of the Hybrid Cloud IaaS model extended combinations are already in practice by giant cloud vendors such as Amazon Web Service and IBM. However, cloud tenants should note that as the model usage increases, the level of complexity, visibility, control and proximity of risk increases as well. Various issues associated to Hybrid, Federated and multi-cloud SLAs (i.e. unplanned downtime, outage, loss of data, Denial of Service (DoS), privacy breach, etc.) have been discussed by (Dhirani, Newe & Nizamani, 2018b). This chapter precisely focuses on the SLA Operations Management (OM) and areas which play a vital role in cloud SLA QoS such as: cloud federation SLAs, Hybrid SLA Management, commercial IT operational management tools, SLA standardization and vendor SLAs, viable solutions followed by the last Section which concludes the chapter.