Developed by SixSq, SlipStream™ is a multi-cloud coordinated provisioning and image factory engine. In other words, it is an engineering Platform as a Service (PaaS) solution to support production deployment in the cloud, as well as development, testing, certification and deployment processes into Infrastructure as a Service (IaaS) cloud environments.

The core of SlipStream™, as well as several connectors, are released under the Apache 2.0 license.

Main Features

In more detail, SlipStream™ currently provides the following features:

• Multi-machine Provisioning System: SlipStream™ allows users to define and execute deployments, based on high-level recipes, independently from the cloud on which the recipes will be applied. Deployments include coordination and orchestration of virtual machines, including ordering and synchronisation of services. Figure 1 shows a screenshot of the SlipStream™ control dashboard of a multi-machine deployment.

  For example, users can deploy, in a single action, on-demand, entire mutli-tier software systems. With SlipStream™, this operation takes a few minutes, often compared to hours if not days using custom solutions.

• Multi-cloud Provisioning: SlipStream™ supports multi-cloud deployments. This means users have the choice of a number of cloud service providers and technologies (public and/or private) when deploying virtual machines, from within the same SlipStream™ service. Furthermore, users can choose to deploy different parts of a deployment on different cloud services or regions, such that redundant and resilient behaviour is reached. Figure 2 shows a screenshot of the SlipStream™ control dashboard of a multi-cloud deployment. SlipStream™ supports different clouds using a connector design. The available connectors are listed in Table 1.

  To illustrate, users can deploy applications (e.g. computing clusters, load balanced applications, multi-media processing pipelines) across several cloud providers, in order to improve data proximity & resilience, while reducing costs.

• Support continuous integration processes with continuous deployment: SlipStream™ encourages users to parameterise image creation and deployment recipes, such that key parameters (e.g. software version, package location, dependencies, inter-relationships) can be provided at runtime. This means that it is easy to integrate SlipStream™ with continuous integration servers to provide a full deployment chain.

  For instance, users can provision complete server-side systems in a single action in order to test and certify a specific version of the system.

• Independence from specific IaaS interfaces and hypervisors: the SlipStream™ recipes are independent of specific IaaS interfaces, therefore avoiding vendor lock-in, allowing you to focus on configuration and deployment, instead of the specifics of each IaaS.

  For example, users can seamlessly apply the same image construction recipes and multi-machine deployments to different cloud providers, yielding the same results every time.

• Community sharing platform: the SlipStream™ data model permits users to decide to share their image and deployment recipes with other users, thus contributing to a community effort, something dear to agile and DevOps principles.

Supported Clouds

SlipStream™ communicates to IaaS clouds services via a connector architecture. A growing number of connectors are available. The connectors talking to open source IaaS implementations are released under the same open source license as the SlipStream™ core, while connectors to proprietary solutions are closed source and available via a commercial license.

The following is the latest update on availability of cloud connectors:

• StratusLab (currently including OpenNebula)
• OpenStack
• Abiquo
• CloudSigma
• Amazon EC2
• VMWare vCloud Director (soon)
• OCCI (soon)
• Physical / Fixed
• CloudStack (soon)
• Microsoft IaaS Azure (soon)

Notes:

1. The Abiquo, Amazon EC2 and CloudSigma connectors are released under a SixSq proprietary license.
2. A VMWare vCloud Director connector is under development, such that SlipStream™ will be able to interface to VMWare-based IaaS cloud services.
3. An OCCI compliant connector is also under development, such that SlipStream™ will be able to interface with OCCI compliant cloud services, e.g. the EGI Federated Cloud services.

Coming soon

The following features are currently under development and will be delivered in 2013:

• Auto-scale and elastic behaviour: users will be able to describe elastic rules, such that SlipStream™ can throttle up and down the cloud resources used. For example, a peak in requests could trigger an increase in the number of worker nodes in a cluster, or the number of web servers in a distributed web service. Inversely, idle resources could be decommissioned so that resource utilization and costs are rationalized.
• Accounting and usage records: users will have the ability to visualize their current resource consumption (e.g. VMs, CPUs, RAM, storage, deployments), per configured cloud. Further, usage records will also be available, allowing users to request a consolidated statement, from two given dates, again per configured cloud. Finally, SlipStream™ system administrators will have the ability to visualize aggregated usage consumption and usage records for each user, such that they can better understand overall resource consumption.
• New look-and-feel with dynamic dashboard: a new skin is under development, improving the user experience and browser support.

Easy access to the cloud

SlipStream™ provides simple interfaces (web and command-line client) to interact with cloud environments. All clouds are not equal and several clouds have their own approach to VM lifecycle and configuration. SlipStream™’s connector architecture provides an abstraction layer giving the opportunity to each connector to leverage the advantages, and work around the inconveniences, of each cloud. The result is a consistent behaviour, across clouds. The following are examples of harmonisation SlipStream™ provides:

• Contextualization: even for cloud solutions that do not support contextualisation, SlipStream™ goes the extra mile and hides this by pushing contextualisation into the VM, after locking any SSH username/password remote access.
• Machine image creation: pick a base image, drop in your scripts (e.g. Bash, Python, Ruby, Chef, Puppet) and/or a list of
  packages in SlipStream™ and let it do the image creation heavy lifting for you. This version and automated process can be applied to all supported cloud services, yielding the same consistent results on these clouds.
• Multi-machine deployment: specify at provisioning time: the number of instances you want for each type of VM in your deployment model, parameters (e.g. package version, credentials, IPs of external systems), as well as the cloud service(s) (you can even deploy different parts of your application on different cloud services) and let SlipStream™ worry about provisioning, orchestrating and configuring your applications.
• Extra disk attachment: keeping your base images small and tidy simplifies life but can be constraining. SlipStream™ is able to automatically attach extra disks, such that you can dynamically increase the available disk size to your virtual machine, at provisioning time.

Our cloud connector architecture allows us to quickly develop and integrate with current and future IaaS cloud services. This means that you will be able to easily switch between clouds, therefore avoiding vendor lock-in. If we do not support your cloud, please get in touch, chances are that we might very well be able to add it to our short term roadmap.

DevOps and SlipStream™

SlipStream™ in a nutshell: Automated, on-demand, creation of multi-machine runtime environments. In other words, a DevOps platform that provides a software engineering PaaS solution, as a product for you to install on a local cloud or as a hosted service.

With SlipStream™, once your deployment is automated, you can assess, from any point in your software development chain, the impact of any software or configuration change on your ability to deploy into production. As an analogy, while JUnit provides a framework for test automation at the class level, SlipStream™ does the same at the system level, including parameterised and multi-node deployment scenarios. Furthermore, SlipStream™ allows you to automate off-nominal scenarios, such as data centre power failures or service unavailability, to verify that your system is able to recover and maintain the expected quality of service your customers expect from your software.

SlipStream™ is a simple to use, yet powerful, application built on top of public clouds or your own infrastructure. Evaluation of SlipStream™ takes a few minutes, using our hosted secure on-line service (SaaS), tutorials and documentation.

The initial idea

How stressful are your production releases? How much confidence do you have in each new release you push to production? These are some of the problems that motivated us in building SlipStream™.

With SlipStream™ we can build on-the-fly runtime or test environments, close to production deployments, which allow us to assert that our good ideas and bug fixes can be deployed into production to deliver value to our users. SlipStream™ works on a variety of IaaS clouds, so that we don’t have to learn the detail of each cloud every time we need to deploy our systems there. This also allows us to share knowledge we have amassed in integrating and testing systems over the years in the form of patterns, machine image creation recipes and multi-machine deployments, configuration and orchestration. You can do the same!

We think SlipStream™ fills this gap. If you see a missing feature, please let us know. We’ll be more than happy to explore adding it to the system.

For techies… How does it work?

SlipStream™ provides two main features:

1. Image creation factory
2. Deployment orchestration and coordination

The image creation factory allows users to create new images automatically by applying a high-level recipe (e.g. scripts, packages) on a base image. The result is a new machine image, built from version controlled elements, which is then simple to configure manage, reproduce, update and maintain.

The deployment feature allows several virtual machines to be deployed together, where SlipStream™ takes care of orchestrating their instantiation and coordinates exchange of configuration information. The tutorial we have prepared, which is available under the documentation link from any SlipStream™ web page, shows how to build a simple client/server deployment where a client contacts a web server to retrieve a simple document. The example shows that the client needs hostname and port of the web server it needs to contact, as well as an indication that the web server is ready to accept connections. For this, SlipStream™ takes advantage of cloud contextualisation (when available), where it executes a script at the end of the boot sequence. This script executes the user defined logic, including synchronisation using a simple messaging feature. The following diagram shows how the orchestrator (a special VM responsible for the execution of the deployment) instantiates two more VMs (Test Client and Apache server). Three messages are then exchanged between the VMs:

• IP address of the web server
• Port on which the web server listens for connections
• A ready state, telling the client that the web server is configured and ready to accept connections

In order to use these messages, scripts are defined by the user in SlipStream™. For more details, please refer to the online tutorial and reference manual.

Cloud Agnostic

SlipStream™ can be configured to talk to several different cloud instances. This means that you can create image recipes and deployments that will yield identical results on different clouds. This is useful when dealing with different clouds running in different locations, or when using different clouds for the different phases of a development life-cycle. The following diagram shows how different users can take advantage of different cloud services, while maintaining their investment in terms of image creation recipes and deployment models.

The figure above also shows how different user profiles can interact with SlipStream™. Here an advanced user will be creating new deployments and images by creating these models in SlipStream™ and interacting directly with the VMs to debug and test the deployments and images. Once in place, standard users can then self-provision VMs and parameterise pre-configured systems, in the click of a button.

For more details, please read the reference documentation and the tutorial, included with our SlipStream™ hosted service.

Try SlipStream™ For Free

Give SlipStream™ a try on our hosted service. The hosted service is currently free of charge.

Creating a SlipStream™ an account takes a few seconds. Once logged-in, you have a number of cloud services to choose from, so simply pick the one you fancy, or several, configure your account accordingly and you’re good to go. We then suggest you start by taking a look at the tutorial.

Software Bundles and Products

SixSq delivers certified SlipStream™ bundles, containing software in the form of binary packages, documentation, started kit and examples. The bundles can be created ‘à la carte’, including open sourced and proprietary connectors, as well as custom image and deployment recipes and scripts. The bundles also include, maintenance and support so that you stay up to date with our fast release cycles.

We also deliver SlipStream™ as in a hardware turn key solution, called emCloud. You can also rely on our SlipStream™ hosted service, if you prefer the piece of mind of a SaaS delivery model.

FAQ

Q1: I don’t know cloud computing, will I be able to use SlipStream™?

A1: Sure. We’ve designed SlipStream™ not to expose cloud-isms in the interface. Cloud computing offers many rich features that are very interesting when deploying runtime systems, but the defaults we’ve put in place will hide most of these details so that you can get going quickly without facing a steep learning curve.

Q2: I know cloud computing, will I be able to do more with SlipStream™?

A2: Absolutely. SlipStream™ does not expose all the richness of each underlying IaaS cloud service it supports. However, it allows you to automate routine operations, such as machine image creation and multi-machine / complex deployments, with a version control system keeping track of your changes. We are leveraging more and more cloud features at the level of SlipStream™, such that you can leverage these features in your deployments.

Q3: What’s the big deal with automating deployments? I’ve got a few scripts that do the trick nicely.

A3: Great. But chances are that you are the only one using these scripts, or others find it difficult to use them. In addition, as your software, its configuration and dependencies change and evolve over time, you have to adapt and change these scripts in a controlled manner and version control them. Furthermore, you might end-up with mixed logic between the steps required to install, configure and start your services and components, from the logic required to bring up your services in the right order. SlipStream™ provides a clear and clean separation of concerns between the installation and configuration of individual services, from the synchronisation logic between services running inside machines. Finally, you should be able to reuse a lot of your current scripts in SlipStream™ and since SlipStream™ has a web interface, others in your team will be able to leverage your efforts by creating private deployments at the click of a button. Finally, your ability to deploy your applications will be independent of the target cloud.

Q4: Can I reuse the same deployment to test different aspects of my system (e.g. functional tests, performance)?

A4: Yes. Each node in a deployment can be instantiated any number of times per deployment. This mean that, for example, you could have a client test that could perform functional tests on its own on a regular basis, while from time to time, you might want to perform the same test with say 100 or 1000 clients running in parallel. You could also control the number of parallel threads/processes your client test uses to contact the server and record access time.

Q5: Can I reuse existing virtual machines in SlipStream™?

A5: Yes. Depending on the cloud back-end you are using, you can create base images that point to existing virtual machines. See the reference manual for further details.