Posts Tagged 'Bare Metal Server'

February 5, 2016

Enable SSD caching on Bare Metal Server for 10X IOPS Improvements

Have you ever wondered how you could leverage the benefits of an SSD at the cost of cheap SATA hard drives?

SSDs provide extremely high IOPS for read and writes and are really tempting for creating volumes, which are IOPS centric. However, because SSD prices are significantly higher than SATA drives, IT managers are at a crossroad and must decide whether to go for SSDs and burn a fortune on them or stay with SATA drives.

But there is a way to use SATA drives and experience SSD performance using some intelligent caching techniques. If you have the right PCI RAID card installed on bare metal servers, you can leverage certain SSD caching feature benefits.

Make sure when configuring a bare metal server, which has sufficient drives bays (8+ the least), to have a LSI (AVAGO) MegaRAID card as the chosen RAID card. You can select the appropriate RAID configuration for OS and other workload data during the order process itself so that the RAIDs come preconfigured with them. As an additional resource for high speed cache device, consider ordering at least two or more SSDs. You can add this to your server even after deployment. These drives are the SSD caching drives that can be used to improve the overall performance of the cheap SATA drives from which one has carved out the volume. 

Install MSM for Easy Management of the RAID Card

Once the server is deployed, consider installing AVAGO MegaRAID Storage Manager (MSM) for the OS that has been installed in the server. (You can also perform a remote management of the RAID controller from a local machine by providing the IP of the server where the controller is installed).

Users can directly download MegaRAID Store Manager from the AVAGO website for the installed card in the machine. For the most popular MegaRAID SAS 9361-8i card download the MSM from the AVAGO website here.

How to Create CacheCade - SSD Caching Volumes and Attach to the Volume Drives

Follow these three steps to improve the IOPS on the existing Volumes on the bare metal server.

Step 1: Creating CacheCade Volumes

Once SSDs are deployed on bare metal servers and Regular Volumes are created, users can create a CacheCade volumes to perform SSD Caching. This can be easily achieved by right clicking AVAGO Controller and selecting the Create Cachecade – SSD Caching option.

Create Cachecade

Step 2: Choosing the right RAID Level and Write Policy for CacheCade Volumes

It is recommended to use a RAID 1 SSD Cache Cade Volume. This will eliminate a single point of failure at the SSD device level. This can be done by selecting available SSDs on the system and choosing RAID 1 as the RAID level. Click Add to add all available disks and Create Drive Group. Also, be sure to select Write Back as the Write Policy for increased IO performance for both Read and Writes to a Volume that needs to be cached. 

RAID Level and Write Policy for CacheCade Volumes

Step 3: Enabling SSD Caching For Volumes

If the Virtual Drives were created without SSD caching enabled, then this is the right time to enable them as shown below—selectively enable or disable set of Virtual drives which needs SSD caching.

Right click on the volume and select Enable SSD Caching.

Enable SSD Caching

Performance Comparison

We tried a simple comparison here on a 3.6TB RAID 50 (3 Drive with 2 Spans) volume with and without SSD caching using IOmeter tool (available here). The workload was a 50/50 (Read/Write) 4kb Pure Random IO workload subjected for about an hour on the volumes. 

Without SSD Caching – IOPS 970

Without SSD Caching IOPS 970

With SSD Caching – IOPS 9000 (10X Improvement)

With SSD Caching IOPS 9000 (10X Improvement)

The result shows a 10X IOPS and workload dependent benefit. Results also show how repeatable the Read/Writes are happening with the same LBA.

This could certainly help a database application or IO centric workloads, which are hungry for IOPS, get an instant boost in performance. Try this today at Softlayer, and see the difference!!

-Subramanian 

 

August 17, 2015

ImageNet Machine-Vision Competitors to Receive GPU-Enabled Bare Metal Cloud Servers from SoftLayer and NVIDIA

For the first time in the history of the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), this year’s qualifying participants will receive free use of bare metal cloud servers equipped with two NVIDIA Tesla K80 dual-GPU accelerators, provided by IBM Cloud and NVIDIA.

Kicking off last Friday, the ILSVRC is an annual object-detection and image-classification competition intended to advance the fields of machine learning and pattern recognition. It’s hosted by the University of North Carolina (UNC), Stanford University, and the University of Michigan.

Over the next three months, teams from around the world will compete to detect, locate, and classify patterns within a huge set of images taken from Internet sources that are tagged with metadata by human volunteers. The overall goal is to develop the most accurate image recognition algorithms with the lowest percentage of classification errors. To read more about the competition, visit NVIDIA’s recent Parallel Forall blog post and the ILSVRC 2015 home page.



Examples of ImageNet images demonstrating classification with localization.

The combination of SoftLayer servers and Tesla K80 GPUs gives teams the most powerful supercomputing cloud servers available in the marketplace today. To give you a quick overview of the specs, each bare metal cloud server comes with:

  • Two NVIDIA Tesla K80 GPU Accelerators
  • Dual Intel Xeon E5-2690 CPUs
  • 128GB RAM, and
  • Two 1TB SATA HDD/RAID 0.

By offering these cloud resources to ILSVRC teams, we’re helping pave the way for advances in the fields of machine learning and deep learning. We’re looking forward to seeing how these teams leverage our powerful, scalable, and secure cloud platform to develop innovative new methods for training deep neural networks.

Our support of this year’s ILSVRC adds to IBM’s rich legacy of providing innovative resources in the machine learning space, including IBM Watson and other software and services. ILSVRC teams are welcome to leverage third-party resources in their approaches, including the IBM Watson Visual Recognition Service, available on IBM Bluemix, and AlchemyVision from AlchemyAPI, an IBM Company.

If you’re interested in joining the competition and getting complimentary access to SoftLayer cloud servers with NVIDIA Tesla K80 GPUs, go to the ILSVRC 2015 home page and register your team. Once accepted into the competition, team leaders will be provided with access methods and credentials by NVIDIA and IBM.

And stay tuned for competition highlights as the ILSVRC continues over the next three months. Winners will be announced in November. Best of luck to all the competitors!

More About IBM Cloud Resources
While IBM Cloud is offering free resources to qualifying ILSVRC participants, the same GPU-enabled bare metal servers are also available to all of our customers in any of IBM Cloud’s SoftLayer data centers. These resources — along with SoftLayer’s high bandwidth, low-latency network, high-performance storage, and data ingestion options like Aspera, Direct Link, and data transfer service — make IBM Cloud the ideal choice for machine-learning deployments in the cloud. To learn more, visit https://www.softlayer.com/gpu.

-Betsy

Categories: 
March 18, 2015

SoftLayer, Bluemix and OpenStack: A Powerful Combination

Building and deploying applications on SoftLayer with Bluemix, IBM’s Platform as a Service (PaaS), just got a whole lot more powerful. At IBM’s Interconnect, we announced a beta service for deploying OpenStack-based virtual servers within Bluemix. Obviously, the new service is exciting because it brings together the scalable, secure, high-performance infrastructure from SoftLayer with the open, standards-based cloud management platform of OpenStack. But making the new service available via Bluemix presents a particularly unique set of opportunities.

Now Bluemix developers can deploy OpenStack-based virtual servers on SoftLayer or their own private OpenStack cloud in a consistent, developer-friendly manner. Without changing your code, your configuration, or your deployment method, you can launch your application to a local OpenStack cloud on your premises, a private OpenStack cloud you have deployed on SoftLayer bare metal servers, or to SoftLayer virtual servers within Bluemix. For instance, you could instantly fire up a few OpenStack-based virtual servers on SoftLayer to test out your new application. After you have impressed your clients and fully tested everything, you could deploy that application to a local OpenStack cloud in your own data center ̶all from within Bluemix. With Bluemix providing the ability to deploy applications across cloud deployment models, developers can create an infrastructure configuration once and deploy consistently, regardless of the stage of their application development life cycle.

OpenStack-based virtual servers on SoftLayer enable you to manage all of your virtual servers through standard OpenStack APIs and user interfaces, and leverage the tooling, knowledge and process you or your organization have already built out. So the choice is yours: you may fully manage your virtual servers directly from within the Bluemix user interface or choose standard OpenStack interface options such as the Horizon management portal, the OpenStack API or the OpenStack command line interface. For clients who are looking for enterprise-class infrastructure as a service but also wish to avoid getting locked in a vendor’s proprietary interface, our new OpenStack standard access provides clients a new choice.

Providing OpenStack-based virtual servers is just one more (albeit major) step toward our goal of providing even more OpenStack integration with SoftLayer services. For clients looking for enterprise-class Infrastructure as a Service (IaaS) available globally and accessible via standard OpenStack interfaces, OpenStack-based virtual servers on SoftLayer provide just what they are looking for.

The beta is open now for you to test deploying and running servers on the new SoftLayer OpenStack public cloud service through Bluemix. You can sign up for a Bluemix 30-day free trial.

- @marcalanjones

August 26, 2014

Bare Metal Power. By the Hour.

Think quickly. You hear that your new app will be featured on the front page of TechCrunch in less than two hours. Because it’s a resource-intensive application you know that a flood of new users will bog down its current cloud infrastructure and you’ll need to scale out.

What do you do? Choose virtual servers to guarantee quick deployment and more flexibility? Opt for bare metal servers to deliver the best user experience (while crossing your fingers that the servers are online in time for the flood of traffic)? In times like these, you shouldn’t have to choose between flexibility and power.

You need hourly bare metal servers.

We’ve streamlined the deployment of four of our most popular bare metal configurations, and with that speed, we’re able to offer them with hourly billing! With the hardware pre-configured, you tell us where you want the server to be provisioned—Dallas, San Jose, Washington D.C., London, Toronto, Amsterdam, Singapore, and Hong Kong—and which operating system you’d like us to install— CentOS, Red Hat, FreeBSD, or Ubuntu. And in less than 30 minutes, your server will be online, fully integrated with your other SoftLayer servers and services, and ready for you.

Use the server for as long as you need it. Spin it down when you’re done. Pay for the hours you had it on your account. It’s that easy. No virtualization. No noisy neighbors. Just your computing-intensive workload, the hardware configuration you need, and a phobia-proof commitment.

Why you need hourly bare metal servers in your cloud life?

  • Processing Power: You have short-term workloads that require significant amounts of processing power. To get the same performance from virtual servers, you might have to provision twice as many nodes or run them for twice as long.
    • Example: a business intelligence ELT (Extract/Load/Transform) application.
  • Schedule-based Workloads: You have a number of applications that require compute and storage resources on a set schedule (i.e., once every month), and you don’t want to deploy (and pay for) high-end machines that will sit idle at all other times.
    • Example: payroll processing or claims payment processing.
  • Performance Testing: Certify or validate how an application performs on a specific hardware configuration.
    • Example: Software or mobile application companies can validate performance on specific hardware platforms.

With bare metal performance available on demand and on hourly terms, you don’t have to compromise performance for flexibility. When TechCrunch comes calling, you have peace of mind that your app’s success and popularity won’t bring it down.

-RJ

Subscribe to bare-metal-server