Welcome!

@CloudExpo Authors: Zakia Bouachraoui, Elizabeth White, Liz McMillan, Pat Romanski, Roger Strukhoff

Related Topics: @CloudExpo, Microservices Expo, Containers Expo Blog

@CloudExpo: Article

Keys to the PaaSing Game: Multi-Tenancy

There are two primary approaches to multi-tenancy in PaaS

Whether SaaS, IaaS or PaaS, one of the central concepts of all layers of cloud computing is multi-tenancy. If there is no shared resource in a deployment, it's difficult to justify calling that deployment "cloud."

Even NIST makes it more or less official within its formal definition of cloud computing, which reads, in part:

Essential Characteristics: Resource pooling. The provider's computing resources are pooled to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand.

In most SaaS offerings, the multi-tenancy is manifold - the servers, the application code, the database and even the individual tables or pages within the database might be shared among different customers and users of the system.

In IaaS, multi-tenancy is implemented via virtualization technology: a hypervisor allocates and manages a number of complete virtual machines on a particular physical computing resource.

But what about PaaS?

There are two primary approaches to multi-tenancy in PaaS: one simply relies upon the IaaS multi-tenancy, which I call Server PaaS. The other looks more like SaaS multi-tenancy. I call that one Resource PaaS.

Server PaaS is essentially an automated deployment and management system. Although there are managed services providers who can manually set up your application development and deployment environments on cloud servers, and perhaps even automate part of it, this is not Platform-as-a-Service. It is just managed services.

If, in contrast, the developer can directly manage the environments via a user interface that offers operations at a high level (even if that user interface is at the command line), then it is actually a service. Examples of Server PaaS include RightScale, Standing Cloud and EngineYard.

Resource PaaS provides an abstract "container" for an application that allows it to share computing resources in a granular way with other such applications. It eliminates the concept of "servers" in favor of functional resources. The extent to which the application container looks similar to a traditional application deployment depends on the particular service (and that's the subject of a whole other article), but in no case is there "root access." Examples of Resource PaaS include Force.com, Google AppEngine, and Heroku.

There are a number of advantages to Resource PaaS. First, application scaling is granular and fast. As the application receives more requests, or starts performing more work, the required resources are available immediately (within limits, of course). Cost calculations are also granular, so you pay only for the computing resources you use. Another advantage mirrors that of SaaS: a developer does not have to think about or manage servers (including servers that fail), or backups, setup, configuration, etc.

But this granularity and abstraction comes with a cost, and that cost is loss of control. Anything that is shared with other users cannot also be configured arbitrarily by each user. If there is a configuration detail that can be changed by individual users, then the software and systems implementing that detail for the user must be separate. Again, each service has its own place on this continuum. For example, Google AppEngine uses the BigTable database, which is shared by everyone, while Heroku allows each customer to set up a separate NoSQL or relational database.

With a Resource PaaS, there is also a loss of control over where the application runs. The provider controls the computing resources and therefore it behaves as a single point of failure (even though there may be redundancies at lower levels). Typically, there is also no way to deploy your application as a "hybrid," where some of the computing resources are owned and others are shared. (CloudFoundry deployments could in the future be an exception to this.)

Because each Resource PaaS is a unique application environment, application code generally must be developed for that particular PaaS and is then somewhat locked in. In the extreme, some PaaS offerings even have a proprietary programming language and can't ever be ported to other environments.

Finally, a Resource PaaS has greater vulnerability to security breaches. This is due simply to that fact that there are more shared resources, so there are more places for bugs in the operating software to be exploited, or to accidentally expose data to "neighbors." Also, hypervisor technology is widely used and has been subjected to a fair amount of security scrutiny, so its isolation of data is relatively well tested. A typical PaaS, on the other hand, does not have as much breadth of usage and so is not subject to the same scrutiny.

The benefits and disadvantages of a Server PaaS are nearly the mirror image of those of a Resource PaaS. Data isolation relies on proven and secure hypervisor technology. The application environment is usually consistent with more traditional deployment approaches, and the application code can be built for portability. If the PaaS supports it, production deployments can be moved, spread across multiple data centers or providers, or organized into a hybrid. Configuration details of the technology stack, and in some cases even the operating system, are generally visible to the developer.

On the flipside, Server PaaS cannot offer the same level of scaling and cost granularity as Resource PaaS. Generally the unit of resource is a server-hour, and adding new resources can take several minutes or longer. Automatically scaling these resources is less accurate because it relies on secondary measures of resource requirements (e.g., CPU load).

A good Server PaaS automates the server management aspects of the application, including not just initial deployment but the production life cycle of the application as well. Done right, it can be nearly as easy to manage as a Resource PaaS, although it does still require incremental awareness.

Which is better? Clearly the answer depends on both current and future needs. If control, flexibility, security, and portability are important to you, then Server PaaS has many advantages. If ease of deployment and management and/or rapid and efficient scaling are crucial, Resource PaaS probably wins. My only generic advice is that for any new applications you build, minimize the dependencies on a particular PaaS or type of PaaS - because you never know when things will change.

More Stories By Dave Jilk

Dave Jilk has an extensive business and technical background in both the software industry and the Internet. He currently serves as CEO of Standing Cloud, Inc., a Boulder-based provider of cloud-based application management solutions that he cofounded in 2009.

Dave is a serial software entrepreneur who also founded Wideforce Systems, a service similar to and pre-dating Amazon Mechanical Turk; and eCortex, a University of Colorado licensee that builds neural network brain models for defense and intelligence research programs. He was also CEO of Xaffire, Inc., a developer of web application management software; an Associate Partner at SOFTBANK Venture Capital (now Mobius); and CEO of GO Software, Inc.

Dave earned a Bachelor of Science degree in Computer Science from the Massachusetts Institute of Technology.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


CloudEXPO Stories
The precious oil is extracted from the seeds of prickly pear cactus plant. After taking out the seeds from the fruits, they are adequately dried and then cold pressed to obtain the oil. Indeed, the prickly seed oil is quite expensive. Well, that is understandable when you consider the fact that the seeds are really tiny and each seed contain only about 5% of oil in it at most, plus the seeds are usually handpicked from the fruits. This means it will take tons of these seeds to produce just one bottle of the oil for commercial purpose. But from its medical properties to its culinary importance, skin lightening, moisturizing, and protection abilities, down to its extraordinary hair care properties, prickly seed oil has got lots of excellent rewards for anyone who pays the price.
The platform combines the strengths of Singtel's extensive, intelligent network capabilities with Microsoft's cloud expertise to create a unique solution that sets new standards for IoT applications," said Mr Diomedes Kastanis, Head of IoT at Singtel. "Our solution provides speed, transparency and flexibility, paving the way for a more pervasive use of IoT to accelerate enterprises' digitalisation efforts. AI-powered intelligent connectivity over Microsoft Azure will be the fastest connected path for IoT innovators to scale globally, and the smartest path to cross-device synergy in an instrumented, connected world.
There are many examples of disruption in consumer space – Uber disrupting the cab industry, Airbnb disrupting the hospitality industry and so on; but have you wondered who is disrupting support and operations? AISERA helps make businesses and customers successful by offering consumer-like user experience for support and operations. We have built the world’s first AI-driven IT / HR / Cloud / Customer Support and Operations solution.
ScaleMP is presenting at CloudEXPO 2019, held June 24-26 in Santa Clara, and we’d love to see you there. At the conference, we’ll demonstrate how ScaleMP is solving one of the most vexing challenges for cloud — memory cost and limit of scale — and how our innovative vSMP MemoryONE solution provides affordable larger server memory for the private and public cloud. Please visit us at Booth No. 519 to connect with our experts and learn more about vSMP MemoryONE and how it is already serving some of the world’s largest data centers. Click here to schedule a meeting with our experts and executives.
Darktrace is the world's leading AI company for cyber security. Created by mathematicians from the University of Cambridge, Darktrace's Enterprise Immune System is the first non-consumer application of machine learning to work at scale, across all network types, from physical, virtualized, and cloud, through to IoT and industrial control systems. Installed as a self-configuring cyber defense platform, Darktrace continuously learns what is ‘normal' for all devices and users, updating its understanding as the environment changes.