This week, I was in beautiful Melbourne, Australia for IBM Systems Technical University.

PowerAI overview and Cognitive Solutions on POWER

Anand Subramaniam, IBM Technical Specialist, presented this session on PowerAI. IBM packaged a collection of Machine Learning libraries, optimized them for POWER8 chip-set, and made this entire package freely available for download as "PowerAI".

(If you don't have access to a POWER8 server, you can try out the "Anaconda" package using Python for x86 laptop or desktop using Windows, Mac or Linux. See Jason Brownlee's [Your First Machine Learning Project in Python: Step by Step] for instructions.)

IBM also is working on a priced value-add collection called "PowerAI Vision"

Hadoop Infrastructure solutions and Point-of-View

Alexis Giral, IBM Executive Storage Architect, presented the benefits of IBM Spectrum Scale using a simple example. Supposed you are gathering 40TB of sensor readings per day. How many TB of storage would you need to hold 2 years worth of data?

Traditionally, HDFS maintains three copies of the data. A recently added feature "HDFS-EC" provides erasure coding to reduce the overall storage requirements. Giral showed this chart:

Method	Scheme	Factor	TB required
Original Source	No protection	1.0	27,740
HDFS	Triplication	3.0	83,320
HDFS-EC	5+4 Erasure Coding	1.8	49,932
Spectrum Scale ESS	8+3 Erasure Coding	1.3	36,894

And this is assuming all the data is hot. If you decide to keep only 30 percent hot, perhaps the most recent eight months, and the other 70 percent on colder storage, you may reduce your storage requirement costs even further.

IBM Cloud Object Storage - Redefining backup infrastructure

Maciej "Mac" Lasota, presented the use of IBM Cloud Object Storage as a backup repository. While IBM Spectrum Protect is the preferred choice, IBM COS also works well with Commvault and NetBackup.

He listed some of the challenges that companies have with backups to tape, and how IBM COS addresses these challenges.

(While IBM COS is three to four times more expensive than tape, it is a luxury many clients can now afford!)

He wrapped up the session showing five different deployments that he worked on for clients.

New Generation of Storage Tiering: Simpler Management, Lower Costs, and Improved Performance

With ever changing amounts of storage, it is hard to find metrics that are consistent year to year. Fortunately, we found I/O density as the metric to focus my efforts, armed with real data from Intelligent Information Lifecycle Management (IILM) studies done at various clients. From that, I was able to talk about storage tiering on three fronts:

• IBM Easy Tier on DS8000 and Spectrum Virtualize to provide tiering within a system.
   
• IBM Virtual Storage Center (VSC) to provide tiering between systems in a data center.
   
• IBM Spectrum Scale, Spectrum Archive and IBM Cloud Object Storage System to provide global tiering across multiple locations, and across flash, disk, tape and cloud resources.

   

Spectrum Scale for Volume, File and Object Storage

Slides are available on IBM Expert
Network on [Slideshare.net]

IBM Spectrum Scale was formerly called GPFS and has been around since 1998. I am glad it was renamed, as GPFS suffered from "guilt by association" with other file systems, AFS, DFS, XFS, ZFS, and so on.

Spectrum Scale does so much more, supports volume, file and object level access, supports POSIX standards for Windows, AIX and Linux, support Hadoop and Spark with 100 percent compatible HDFS Transparency Connector, support NFS, SMB and iSCSI protocols, as well as OpenStack Swift and Amazon S3 object based access.

Initially designed for video streaming and High Performance Computing (HPC), IBM has extended its reach to work in a variety of workloads across different industries. More than 5,000 production systems are running at client locations.

Beating Ransomware! A deep exploration of threat vectors for applications and storage

Andrew Greenfield, IBM Global Engineer for Spectrum Storage, presented on the threat of ransomware. In addition to being an expert in various storage, he also is an expert in security.

If you think security is just setting up your network firewalls and turning on data-at-rest encryption on your storage, you are sadly mistaken. Many of the treat vectors come from the inside, disgruntled employees or temporary contractors who plant viruses, bombs and worms that may not activate until long after they leave.

There are now products called security information and event management (SIEM) that provide real-time analysis of security alerts generated by network hardware and applications. Two that Andrew was familiar with were IBM Qradar and Varonis. These identify standard and abnormal behavior patterns among users.

Andrew feels products like Splunk do a great job to collect information, but don't do the analysis that Qradar or Varonis do.

I was very pleased with this conference. This was a concentrated 3-day event, but everyone I talked to was happy with the format, and felt their time spent worthwhile!

technorati tags: IBM, #IBMtechU, #IBMstu2017, Anand Subramaniam, PowerAI, PowerAI Vision, Anaconda, Machine Learning, Alexis Giral, Hadoop, HDFS, HDFS-EC, Erasure Coding, Spectrum Scale, Elastic Storage Server, ESS, Maciej Lasota, Mac Lasota, IBM Cloud Object Storage, IBM COS, Spectrum Protect, Commvault, NetBackup, Easy Tier, Virtual Storage Center, VSC, Spectrum Archive, GPFS, POSIX, NFS, SMB, iSCSI, OpenStack Swift, Amazon S3, High Performance Computing, HPC, Andrew Greenfield, IBM Qradar, Varonis, SIEM, Splunk

This week, I was in beautiful Melbourne, Australia for IBM Systems Technical University. On Wednesday evening, we had a poster session.

(I have so many photos that I will split this post up into topics. This post will focus on IBM Z systems, see my other posts for storage and IBM Power systems.)

Topics can be anything that is of interest to your peers and colleagues. It can be research-related, a specific solution you implemented or an interesting customer case you want to share.

Linux Scalability at a Small Scale (or, An Adventure In Minimalist Multitudinousness)

Vic Cross, IBM Senior Systems Engineer, used the Ganglia Monitor System to generate traffic and measure 1,680 Linux guests on a single IBM Z mainframe LPAR with only 16GB of memory! His poster consisted of 18 pages of material, a mix of traditional presentation slides, screen shots of web pages, and densely detailed performance results.

Ganglia is a scalable distributed monitoring system for high-performance computing systems such as clusters and Grids. It is based on a hierarchical design targeted at federations of clusters. It leverages widely used technologies such as XML for data representation, XDR for compact, portable data transport, and RRDtool for data storage and visualization. It uses carefully engineered data structures and algorithms to achieve very low per-node overheads and high concurrency. The implementation is robust, has been ported to an extensive set of operating systems and processor architectures, and is currently in use on thousands of clusters around the world. It has been used to link clusters across university campuses and around the world and can scale to handle clusters with 2000 nodes. Learn more at [http://ganglia.sourceforge.net/]

Spectrum Scale 2 site cluster

Antony Steel, IBM Senior Consulting IT Specialist, presents an option to configure a 2 site GPFS (Spectrum Scale) "almost active-active" cluster when a 3rd site is not available. This option will require simple administrative tasks to make DR filesystem available should Production site fail. Spectrum Scale runs on IBM Z, IBM Power and x86 servers.

The poster used 13 traditional landscape slides, printed on what appears to be A4 paper. A4 is 297 mm wide, so three side by side exceeds the 841 mm width of the poster foam board. These were arranged with a title slide on top, and then 12 content slides in four rows of three.

While I was glad that someone else had a QR code on their poster, the placement was way at the top, and difficult for anyone to actually scan it. I thought of this, and had mine at waist level in the middle right side of my poster.

Life is better with Linux

I couldn't resist taking a photo of the back of this guy's tee-shirt, which says "Life is better with Linux"

In effect, tee-shirts can also be "posters", although that would make for an awkward "poster session" if everyone wore them? Pointing at your chest would be weird, and pointing to your back would be near impossible!

In 1999-2001, I helped the port of Linux to IBM S/390 mainframe chip-set architecture by testing and debugging the disk and tape device drivers. I was the first to install Linux on an IBM mainframe in Tucson, AZ!

I would then go on to work with SAN Volume Controller, Tivoli Storage Manager (now called Spectrum Protect), Tivoli Storage Productivity Center (now called Spectrum Control), and the General Parallel File System (GPFS, now called Spectrum Scale). All of these run on Linux!

I would become the "Linux storage expert" at conferences like SHARE and GUIDE. While my co-workers in DFSMS and z/OS felt Linux was just a fad, I predicted that Linux was going to be a major force in the IT industry. I was right, not only does Linux run on all of our IBM Z and Power servers, it is the underlying operating system for nearly all of IBM storage devices.

Today, I run Linux directly on my laptop, using a Windows KVM guest image as needed for specific projects or applications.

Erina Araki poses for a photo with one of the attendees, Marco. Erina was the organizer for this poster event, and was my primary contact to answer all of my questions. I think the poster session was a big success!

technorati tags: IBM, #IBMtechU, #IBMstu2017, Ganglia Monitor System, Vic Cross, Anthony Steel, GPFS, QR Code, Linux, S/390, SAN Volume Controller, SVC, Tivoli Storage Manager, TSM, Spectrum Protect,Tivoli Storage Productivity Center, TPC, Spectrum Control, GPFS, Spectrum Scale, DFSMS, z/OS, SHARE Conference, GUIDE Conference, Linux KVM, Erina Araki

This week, I was in beautiful Melbourne, Australia for IBM Systems Technical University. On Wednesday evening, we had a poster session.

(I have so many photos that I will split this post up into topics. This post will focus on posters related to IBM Power systems. See my other posts for storage and IBM Z.)

Ding! IBM i Systems Management redefined with SQL

A poster presentation should trigger question-and-answer sessions, and the exchange of ideas and information regarding your topic.

Scott Forstie, IBM Db2 for i Business Architect, coined the phrase "Scott's Query Language", focused on Data Services for Db2 database on IBM i operating system. His design took several charts, printed in landscape mode, and organized in 3 columns of four charts each. His "title" page was printed twice, and placed on the left and right sides.

Scott explained GROUP_PTF_CURRENCY, LICENSE_EXPIRATION_CHECK and ACTIVE_JOB_INFO. I am not familiar with any of these things, but I enjoyed how passionate Scott was. He even had business cards for people to get more information at: [ibm.biz/Db2foriServices]

IBM Spectrum Scale with Hortonworks Data Platform

Chris Maestas, IBM Global Senior Solutions Architect, IBM and Par Hettinga, IBM Global SDI Enablement Leader, created this poster.

Hortonworks is a leading innovator in the industry, creating, distributing and supporting enterprise-ready open data platforms and modern data applications. They focus on driving innovation in open source communities such as Apache Hadoop, NiFi, and Spark. Their product, Hortonworks Data Platform (HDP), runs on both x86 and Power systems.

The poster design was clean, with basically three enlarged presentation slides. On the top, it explains that Hortonworks now supports IBM Spectrum Scale for storage of files and objects to be analyzed by Hadoop. On the bottom left, it shows how Spectrum Scale eliminates the ingest-and-discard approach used by other HDFS-based systems. On the bottom right, an architecture diagram to build your own "data lake".

Optimizing Power Performance with Affinity Groups – Real World 40Gbit LPM Results / Lessons Learnt

This poster employed a unique 1-6-6 design. Top slide was for title and author: Stephen Diwell, Senior Power Systems Engineer, DXC Technologies

In the middle, the poster had six traditional text-only presentation slides, arranged in two rows of three. LPAR Affinity Groups provides you the ability to give the Hypervisor a hint that you would like this group of LPARs to be located on processor Chips that are closer to each other. Use Affinity Groups to help the Hypervisor place LPARs nearer to the VIO Servers. LPARs that share common resources, like the Fibre Channel and Ethernet adapters within a VIO servers will obtain better performance and adapter throughput the closer they are. The lighting on some of these posters was really poor, and perhaps too dark to read small fonts like this.

At the bottom were performance bar chart results, in three rows of two. I like the use of color for the graphs. For a network job with 8 threads, Stephen achieved a 54% increase in network bandwidth for LPARs communicating on the same Chip to those communicating between Nodes in the E800 frame.

Sundata Power Server Cloud offering

Leave it to the marketing department of a local cloud service provider to turn their poster into an advertising billboard! This one was presented by Kon Kakanis, Managing Director, Sundata Pty Ltd

The Sundata poster encouraged people to move their AIX, IBM i and Linux on POWER workloads to their "PowerCloud" platform. They summarized their advantages into four bullet points:

• Reliable and cost-effective partnership
• Advice, Guidance and Support
• Migration, management and support services
• Located in Sydney and Brisbane

Founded in 1986, Sundata is an Australia-based organization to help their clients transform into the Cloud, select and deploy IT hardware and keep the lights on with ongoing support and managed services. They have over 100 corporations, government departments and schools enjoying a close and ongoing relationship.

The large fonts, simple design, and the cute cat-in-a-cape logo in the lower right corner captured peoples attention!

In between reading posters and talking to everyone, it was good to take a quick look out the floor-to-ceiling windows. At 297 meters, Eureka Tower has some amazing views. Here is one of the Yarra river and Central Business District.

technorati tags: IBM, #IBMtechU, #IBMstu2017, Scott Forstie, Chris Maestas, Par Hettinga, Kon Kakanis, IBM Db2, IBM i, Hortonworks, Hortonworks Data Platform, HDP, Spectrum Scale, Poster Session, IBM Power, Affinity Groups, Sundata, PowerCloud, AIX, Linux on Power, Brisbane, Sydney, Eureka Tower, Yarra River

This week, I was in beautiful Melbourne, Australia for IBM Systems Technical University. On Wednesday evening, we had a poster session. This was the first time I presented a poster session, so I was understandably very excited.

(I have so many photos that I will split this post up into topics. This post will focus on storage posters. See my other posts for IBM Power and Z systems.)

The venue was Eureka Skydeck 89, the top floor of the Eureka Tower. This tower is 297 meters tall (974 feet), and the views it afforded of the city of Melbourne were stunning.

Mo and I arrived early as I was one of the 11 finalists that got selected to present a poster. While it is a hot summer back in Arizona, it is cold here in Australia. I am glad we brought our heavy coats for the brisk 8-minute walk from our hotel, the Crown Promenade, to the Eureka Tower.

Posters are designed to present specific topics in a concise and interactive way to appeal to peers and colleagues at conferences and/or public displays. Everyone would be given an "A0" poster size foam board on which to tape on their poster, 841mm wide, and 1189 mm tall (roughly three feet by four feet).

Understanding Converged and Hyperconverged Systems

My design was simple. I took my summary chart from one of my presentations, and enlarged it to fit the "A0" poster size. I chose my "Pendulum Swings" presentation that explains the history of storage infrastructure, and the rise in interest in Converged and Hyperconverged Infrastructure.

In the early days of IT, storage was internal to its server, over time, storage outgrew its container, and we started having externally attached storage, and benefits like RAID and clustered servers for high availability. Then, SANs, LANs and WANs took the main stage, allowing for greater connectivity and distance.

But now, it seems the pendulum is swinging back with converged and hyperconverged systems. Converged Systems like IBM PureSystems, or VersaStack from IBM and Cisco, provide best-of-breed hardware for servers, storage and networks in a pre-cabled, pre-configured rack. With everything in a single rack, port count and cable distance limits are no longer a major concern.

Hyperconverged Systems, such as IBM Spectrum Scale, IBM Spectrum Accelerate, Nutanix or Simplivity, focus instead on offering commodity servers with internal flash and disk storage. Software-Defined Storage software is then used to glue together multiple units over a LAN infrastruture. With the huge increase in Flash and Disk capacities, a server with internal storage can hold many TB of data.

My poster included a "QR Code" that pointed to a link on BOX so that people could use their smartphones to access all of my presentations.

IBM Spectrum Scale with focus on Active File Management

A poster presents not all the details but the most important information.

Trishali Nayar, IBM AFM/Spectrum Scale Development from Pune India, had a poster on IBM Spectrum Scale with focus on Active File Management (AFM). She had a clean, simple design, basically two presentation slides enlarged to fill the poster size.

Active File Management (AFM) enables sharing of data across clusters, even if the networks are unreliable or have high latency. AFM allows you to create associations between IBM Spectrum Scale™ clusters or between IBM Spectrum Scale clusters and NFS data source. With AFM, you can implement a single name space view across sites around the world making your global name space truly global. You can also duplicate data for disaster recovery purposes without suffering from WAN latencies.

IBM Ubiquity Storage Service for Container Ecosystems

Your audience isn't trying to replicate your solution or case -- they are simply after the basics. Take for example, this poster on IBM's Ubiquity Storage Services.

Ashutosh Mate, IBM WW Senior Solutions Architect, created this poster on storage for Containers. Not to be confused with the Containers used in Spectrum Protect container pools, or the Containers supported by IBM Cloud Object Storage!

The poster had six enlarged presentation slides. Two at the top under "Abstract" covered business need and technology overview. The two in the middle under "Ubiquity Architecture" had a connection diagram and a list of supported environments. The last two under "IBM Vision" covered customer value, use cases, and additional resources.

As people transition from monolithic applications to microservices, IT is shifting from heavy Virtual Machines to lightweight Docker containers.

The Ubiquity project enables persistent storage for the Kubernetes and Docker container frameworks. It is a pluggable framework available for different storage systems. The framework interfaces with the storage systems, using their plugins. Different container frameworks can use Ubiquity concurrently, allowing access to different storage systems.

IBM has support for Spectrum Scale, all of the Spectrum Accelerate offerings (including XIV, FlashSystem A9000/R) and all of the Spectrum Virtualize offerings (including SVC, Storwize and FlashSystem V9000).

Single page handouts as "take-aways" was a nice extra touch.

technorati tags: IBM, #IBMtechU, #IBMstu2017, Melbourne Australia, Poster Session, Ashutosh Mate, Eureka Tower, Eureka Skydeck, Eureka 89, Crown Promenade, Pendulum Swings, Converged Infrastructure, Hyperconverged Infrastructure, HCI, QR Code, Trishali Nayar, Spectrum Scale, Active File Management, AFM, WAN, Ubiquity Storage Services, Docker, Kubernetes, Spectrum Accelerate, Spectrum Virtualize, FlashSystem, XIV, SVC, Storwize