IBM Puffs Up Platform Symphony, LSF On SoftLayer Cloud
IBM entered 2014 hitting the cloud drum hard. Taking a page from tech titans like Google, Microsoft, and Amazon Web Services, Big Blue wants to make it as simple as possible for enterprises to develop and deploy applications.
For workloads that require high-performance hardware and the software to manage workloads on it, IBM has added the Platform Computing Cloud Service to its ever-widening cloud portfolio. The cluster-as-a-service environment targets organizations with compute-intensive analytics and technical computing applications. Depending on the needs of the user, it can be employed as a hybrid cluster for adding capacity during times of peak demand (aka cloud bursting), and also as a stand-alone cluster in the cloud.
The new service is an outgrowth of IBM's acquisition of Toronto-based workload management vendor Platform Computing two years ago, and runs on top of the SoftLayer infrastructure, acquired by IBM last year. The service combines fully integrated workload management, including the Platform LSF grid job scheduler and Platform Symphony Java messaging grid software, spanning on-premise and on-cloud resources with the support of a dedicated cloud operations team.
"This marries the best of dynamic high performance computing capabilities of IBM Platform Computing software with the infrastructure as a service (IaaS) capabilities of SoftLayer," notes senior certified IBM architect Gary Zeien in a recent blog, where he explores what applications will run best on the service. (Hint: it's not just traditional HPC applications.)
For a subset of users for whom performance is critical, typical enterprise clouds that rely on server virtualization won't suffice. These users would like to leverage the benefits of cloud (scalability, elasticity, convenience, and so forth) but they are wary of the performance penalty incurred by virtualization. The IBM Platform Computing Cloud Service addresses these concerns by employing dedicated bare-metal servers. Non-shared physical machines add an extra level security.
The primary use case for the new cloud is the hybrid or burst scenario. Users will be able to access cloud resources on a temporary basis without the need to purchase, configure and maintain new in-house hardware. IBM builds the environment for the customer, installing needed software in the on-premise and off-premise systems. Capacity is expanded as needed by bursting jobs from on-premise to the SoftLayer environment. The idea behind cloud bursting is that it enables organizations to better meet changing business needs and complete projects more quickly.
The IBM Platform Computing Cloud Service includes Platform LSF or Platform Symphony workload management software delivered as a pay-per-use service. Garnered from the Platform Computing acquisition, LSF and Symphony can benefit technical computing, analytics, and service-oriented workloads in the following ways, according to IBM:
- Improved time to results and faster time to market
- Reduced costs
- Simplified workload management
- Enhanced user productivity
- Single source for end-to-end cluster support
Platform LSF provides bursting capabilities and is most often used to boost usage in traditional technical computing environments, where it is common for applications to employ the Message Passing Interface (MPI) for parallelized workloads.
Platform Symphony enables distributed Java applications and analytics to run on a scalable shared grid. Designed to deliver performance benefits in service-oriented environments, Platform Symphony is widely used for SOA workloads in fields such as financial services and life sciences. An Advanced Edition of Platform Symphony includes an Apache Hadoop-compatible MapReduce implementation that has been optimized for speed and reliability. Unlike the open source Hadoop, the popular big data analytics framework, IBM's version has added the capability to re-start failed services automatically.
IBM is mainly targeting the Platform Computing Cloud Service at existing LSF and Symphony customers who already have a cluster and have a need for additional cluster resources. However, the cluster-on-demand solution is also available for prospective Platform clients seeking additional resources who lack the budget or expertise required for an in-house system.
The Platform cluster is sold as a service, including management and support, for a monthly or hourly fee, available now. As is customary with the cloud computing model, customers pay only for what is used, and can scale capacity up and down as needed in an elastic fashion, allowing costs to be shifted from a capital to operational basis. Pricing for the Platform services was not available.