CLOUD COMPUTING

Cloud Computing

The term cloud computing has been tossed around in the computing world with a lot of different meanings. In HPC some researchers use the cloud to run all of their applications, and not have to worry about the cost to purchase, run and maintain their own hardware. Other researchers will use the cloud as a “Cloud Burst.” This can be done in different ways depending on how much resources are needed, and how much a researcher is willing to pay. Find out more about cloud bursting. Finally, some Data Centers will have their own clouds for their users. This allows users to set up their own “space” for their research that can be deleted, changed, or saved for later. It also allows for different Operating Systems to be used for their research. For instance, it’s not trivial to put a Windows Database on an HPC cluster, but it is to do so in the cloud. At Aspen Systems, we have the ability to help you set-up your cloud infrastructure, and match up the hardware with the software to fit your needs.

Open Source Creation of Private and Public Clouds

To help you streamline the management of your HPC systems, Aspen Systems has partnered with OpenStack, a cloud computing operating system that controls large pools of HPC resources throughout a data center. OpenStack provides systems administrators with a dashboard for controlling large pools of networking, computing and storage resources. OpenStack can also be managed and controlled via the OpenStack API for the enablement of complete system integration.


Cloud Computing from Openstack Software Solutions Logo
Openstack Diagram


Take a tour of the new features and enhancements of Mitaka, OpenStack’s thirteenth software release that helps organizations manage cloud infrastructure.

How is OpenStack used in a cloud environment?

The cloud is all about providing computing for end users in a remote environment, where the actual software runs as a service on reliable and scalable servers rather than on each end-user’s computer. Cloud computing can refer to a lot of different things, but typically the industry talks about running different items “as a service”—software, platforms and infrastructure. OpenStack falls into the latter category and is considered Infrastructure as a Service (IaaS). Providing infrastructure means that OpenStack makes it easy for users to quickly add new instance, upon which other cloud components can run. OpenStack works with popular enterprise and open source technologies making it ideal for heterogeneous infrastructure.

“As we examined the various initiatives we’d have underway in our cloud, which would include the intermingling of physical and virtual loads as well as complex network configurations, we knew that the flexibility inherent in OpenStack would prove powerful enough to support our objectives. With OpenStack we have the ability to customize the platform and maintain significant control over our destiny.”

REINHARDT QUELLE, OPERATIONS ARCHITECT, CISCO WEBEX

CloudStack Logo

Open Source Cloud Computing

CloudStack is open source software designed to deploy and manage large networks of virtual machines, as a highly available, highly scalable Infrastructure as a Service (IaaS) cloud computing platform.

CloudStack is used by a number of service providers to offer public cloud services, and by many companies to provide an on-premises (private) cloud offering, or as part of a hybrid cloud solution. Users can manage their cloud with an easy to use Web interface, command line tools, and/or a full-featured RESTful API.


Features and Functionality

Apache CloudStack is a Java-based project that provides a management server and agents (if needed) for hypervisor hosts so that you can run an IaaS cloud. Some, but not all, of the features and functionality provided by CloudStack:

  • Works with hosts running XenServer/XCP, KVM, Hyper-V, and/or VMware ESXi with vSphere
  • Provides a friendly Web-based UI for managing the cloud
  • Provides a native API
  • May provide an Amazon S3/EC2 compatible API (optional)
  • Manages storage for instances running on the hypervisors (primary storage) as well as templates, snapshots, and ISO images (secondary storage)
  • Orchestrates network services from the data link layer (L2) to some application layer (L7) services, such as DHCP, NAT, firewall, VPN, and so on
  • Accounting of network, compute, and storage resources
  • Multi-tenancy/account separation
  • User management

In short, organizations can use Apache CloudStack to deploy a full-featured public or private IaaS cloud.


Hedvig Logo

Hyperscale Storage

Hedvig is a modern, flexible approach to storage. Designed to span both private and public clouds, Hedvig creates a single, programmable data services layer for both traditional and emerging workloads. Hedvig’s platform is powered by patented distributed systems technology, transforming everyday server infrastructure into flexible, modern storage that can be tailored to any workload.

Software-Defined Storage

The Hedvig Distributed Storage Platform provides a single unified solution that lets you tailor a modern, high-performance, elastic storage system built with low-cost commodity hardware to support any application, hypervisor, container, or cloud.






See how Hedvig supports hybrid cloud infrastructures and DevOps with software-defined storage for cloud computing, including Docker, VMware and OpenStack.

Which solution is right for me?

Running HPC applications on the cloud has become more of a commonplace. When funding for hardware comes into play, researchers are beginning to ask what they should do. Buying a new or addition to their cluster, buying time to run on the cloud, and a hybrid of the two are the three most common options. We’ve outlined some of the pros and cons of each option below.

Location Pros Cons
On Premise
  • More overall control
  • You own the hardware and can run on it until it fails.
  • Overall, it can be cheaper than running in the cloud.
  • More control of who runs and who doesn’t run, and how jobs are shared on resources.
  • Local data transfers with long term storage.
  • Easy to install using licenses.
  • You control the security.
  • You control the upgrades.
  • Have to manage the cluster software and hardware.
  • Need to purchase infrastructure.
  • Need to purchase the hardware.
  • Need to keep OS/Software up to date.
  • Limited amount of hardware/resources.
  • You need to worry about the security of your cluster.
Off Premise Cloud
  • Don’t need to worry about Electric bills.
  • No need for infrastructure, power, and cooling.
  • You don’t need a Systems Administrator.
  • Access to multiple types of hardware.
  • You can work on your jobs from anywhere, just the same.
  • Can share information with others in remote locations.
  • Usually backed up.
  • Can get your own environment with control over it.
  • Can cost less if you add infrastructure, power, and Systems Administrator costs. Especially if you run during “cheaper” hours.
  • More resources than needed.
  • Less overall control!
  • Data has to be moved before and after runs.
  • Need to set up software.
  • Expensive to run jobs that fail.
  • Can have issues with using licenses.
  • You have to trust security to the providers.
  • Software updates can cause issues with compatibility.
  • Cloud providers can shut down without or with very little notice.
Hybrid (Cloud Bursting)
  • Best of both worlds
  • Add to resources that you have on premise.
  • Great for deadlines when you don’t have enough resources locally.
  • Software setup to burst to the cloud.
  • Same cons as running on the cloud.
  • Not knowing when you will run on the cloud.

If you are at a University, chances are, you don’t need to use your funding for the facilities or the power. If a location to place your cluster is available, most would do well by housing it on premise. If you need someone to help manage your cluster, Aspen Systems has multiple options for remote systems administration services.

University of Nebraska Logo

The folks at Holland Computing Center at the University of Nebraska-Lincoln ran the numbers for you. They found that it would cost about $4 Million to run on AWS with the same performance as one of Aspen System’s built HPC Clusters. This cluster hardware (with HPC Networking) came in at much lower than $4 Million to run, and they are able to own the hardware. There are times when the cloud makes sense, but having your own machine can come at a lower price.