With public cloud providers poised to become indispensable utility providers, neutrality-related mandates will likely emerge to ensure a level playing field among their customers ("tenants"). We analogize with net neutrality to discuss: (i) what form cloud neutrality might take, (ii) what lessons might the net neutrality debate have to offer, and (iii) in what ways cloud neutrality would be different from (and even more difficult than) net neutrality. We use idealized thought experiments and simple workload case studies to illustrate our points and conclude with a discussion of challenges and future directions. Our paper points to a rich and important area for future work.
A natural but relatively little addressed set of concerns in the emerging public cloud utility comes to the fore when one compares its likely evolution with that of Internet Service Providers (ISPs). During the period of commercialization of the Internet in the 1990s (after the creation of the WWW), the principle of over-engineering reigned and cheap access plans without quotas were commonplace. In the 2000s, the Internet access marketplace matured, ISPs consolidated, and it became apparent that the Internet was (i) prone to congestion by activity that generated little revenue1, and (ii) created a platform for stiff competition for the ISPs' own profitable "managed" video and voice services.
The network neutrality debate began when Comcast throttled BitTorrent activity that was congesting its (broadcast-based CMTS) residential broadband service. Though this violation of application neutrality targeted activity that was largely piracy of copyrighted material, third-party content and service providers viewed it as a dangerous precedent considering that ISPs are themselves competing providers of content (particularly video) and services (e.g., telephony). In 2014, Neflix negotiated payment for "fastlanes" to reach Comcast's subscribers , but this would have basically been a side payment (albeit willing), thus violating origin neutrality. The debate continues: a federal court limited certain of the FCC's neutrality rules in 2014  and the FCC deemed the Internet a utility in 2015  in a move to consolidate neutrality rules.
Somewhat similarly, the public cloud market was also plentiful initially compared to demand. As the business has grown over the last decade, competition has intensified with an increased diversification in the types of offered service-level agreements (SLAs) and associated price-performance trade-offs [4, 30, 8, 19, 13]. It is reasonable to anticipate that issues of fair competition similar to those for ISPs may be on the horizon. As an illustrative example, consider how Amazon Prime's video service and Netflix both use Amazon EC2.
What might neutrality mean in this setting? Do lessons and challenges from net neutrality suffice or does the public cloud raise novel issues? We believe the answer is that it does and this is the context of our paper.
Contributions and Outline: In Section 2.1, we identify several ideas and lessons from the net neutrality debate that offer useful starting points for discussing cloud neutrality. A public cloud is, however, a significantly different resource provider than an ISP. In fact, it is significantly more complicated in some ways, presenting novel concerns that we discuss in Section 2.2. Following this, in Section 3, we take preliminary steps towards exploring these issues using simple thought experiments and empirical case studies. Finally, we conclude in Section 5.
Discussion on public cloud neutrality is relatively scant or preliminary . Interxion, a European provider defines a carrier- and cloud-neutral "colo" data center as follows: "A truly neutral data centre provider is one that is independent of the companies colocating in the data centre, does not compete with them in any way, and offers no packaged services as part of colocation. Customers are free to contract directly with the providers of their choice" . Our interest herein is with public cloud providers that offer more general services than colos (IaaS clouds offering virtualized IT resources or PaaS/SaaS clouds offering even more abstract services), for which the neutrality discussion is much more complex and unclear.
The following insights and guiding principles from the net neutrality debate offer good starting points for our discussion of cloud neutrality.
Resource congestion: If a utility provider can offer adequate resources to satisfy its customers - especially when it is in its initial growing phase e.g. ISPs in the 90s or public clouds till very recently - the neutrality concern is moot. Resource congestion, whether actual or alleged, is the basic context of any neutrality debate.
Neutrality modulo SLAs: Neutrality does not mean that every tenant is treated equally. The fair/equal treatment across tenants may only relate to tenants that have chosen identical SLAs. Therefore, it only ever makes sense to discuss neutrality in the context of specific SLAs. In other words, a key aspect of a provider's resource allocation when assessing its neutrality is how it allocates any discretionary resources (i.e., resources left over after it has met its SLA obligations) among tenants in the same SLA class. This view will play an important point in our case studies in Section 4.
Information limits and preferences: The providers behavior must not be based on "inside information" or "preferences." E.g., Mogul et al.  consider a network bandwidth allocation problem wherein a cloud provider reckons tenant sensitivity to network bandwidth underprovisioning and uses this to allocate bandwidth differentially to improve its profits. One may consider such a cloud provider non-neutral.
Fair competition: Fair competition needs to be upheld between affiliates of the cloud provider and its tenants - whether direct (e.g. recall from Section 1 how Amazon Prime and Neftlix both use Amazon EC2) or indirect. That is, neutrality is related to antitrust.
Dr. Bhuvan Urgaonkar, PhD has over 15 years of experience in the field of Software Engineering and Computers. His work includes research in computer systems software, distributed computing (including systems such as Zookeeper, Redis, Memcached, Cassandra, Kafka), datacenters, cloud computing, storage systems, energy efficiency of computers and datacenters, big data (including systems such as Hadoop, Spark). He serves as an expert / technical consultant with multiple firms helping them (i) understand technical content related to state of the art products in areas such as content distribution, distributed computing, datacenter design, among others and (ii) interpret patents in these areas and connections between them and state of the art products and services. Services are available to law firms, government agencies, schools, firms / corporations, and hospitals.
©Copyright - All Rights Reserved
DO NOT REPRODUCE WITHOUT WRITTEN PERMISSION BY AUTHOR.