Monday Mar 24, 2014

Four High Performance Configurations for SuperCluster and SPARC Servers

When a surfing buddy of mine asked me to look at a banking application that runs on Solaris, I thought he'd been hit on the head by his board one too many times. Solaris is cool. Banking is not. But I looked into it, anyway, and to my surprise, I found the banking app had a certain amount of geek appeal.

If geek appeal is not enough to hold your interest, Mister Hair-on-Fire, the other reason for talking about this banking application is that it helped identify four high performance configurations for Oracle's SuperCluster and SPARC servers that might be useful for other types of applications. So keep reading. Or ...

Go directly to white paper (pdf) that describes the configurations.

What first caught my interest was the idea of a bank operating system. A traditional computer OS manages hardware devices and provides services for application software. A bank headquarters does something very similar. It manages the branches (hardware) and provides services for its operations (applications). Turns out, that's the idea behind Finacle's Core Banking Solution.

Core banking sounds dull as hell, but it's a big deal for banks. It replaces cumbersome end-of-day consolidation between branch banks and HQ. (I almost feel asleep just writing that.) In fact, centralized banks worldwide now mandate the implementation of core banking technology to prevent fraud and meet regulatory requirements.

As a result, Finacle's Core Banking Solution is designed as configurable modules with layered Service-Oriented Architecture (SOA), straight-through processing (STP) capabilities, web-enabled technology, and support for 24 x 7 operations.

But no matter how sophisticated the application, the underlying architecture can limit its performance. Not a problem! Since Finacle 10 is now available on Oracle Solaris, it can be run on the screaming fast Oracle SuperCluster or Oracle’s SPARC T-Series servers. As you might expect, Finacle tested this combination for both batch and OLTP processing and found:

  • Batch results that processed 15% more accounts and 3.2 to 3.7 times the required minimum records per second, all achieved within one third of the specified time, with plenty of CPU resources available to handle further load.
  • OLTP results that exceeded Finacle acceptance criteria with more users and more transactions per second, all with sub-second response times and with considerable CPU resources remaining available.

White Paper: Infosys Finacle Core Banking Solution on Oracle SuperCluster and Oracle’s SPARC T-Series Servers

Roger Bitar provides technical details about the software and hardware layers in this solution, and describes the configurations that obtained the best performance:

  • Configuration for Fastest OLTP Processing on SuperCluster T4-4
  • Configuration for Fastest Batch Processing on SuperCluster T4-4
  • Configuration for Fastest OLTP Processing on SPARC T4-4 Server
  • Configuration for Fastest Batch Processing on SPARC T4-4 Server

About the Photograph

I took this picture of bike race taken in Durango, Colorado, in the Fall of 2012.

- Rick

Follow me on:
Blog | Facebook | Twitter | YouTube

Wednesday Jan 29, 2014

Man Vs. Machine

Man vs machine. Command Line vs GUI. It's not a new debate. In fact, when I was a little boy I watched this Paul Bunyan cartoon with the dismay of a sysadmin watching the increasing popularity of GUIs:

Cartoon: Paul Bunyan's Ax vs The Chain Saw

What Skills Do Sysadmins Need to Manage a Modern Data Center?

Video Interview with Brian Bream

When I wrote technical manuals for Oracle Solaris back in the day, I had the luxury of my very own lab. For instance, while writing the NIS+ books, I was able to discover my own procedures on a small network and, when I needed something larger, I could ask the sysadmins in Sun's bigger labs to try some experiments for me. Little did I know those were the Golden Years of technical writing.

They were also the Command Line Years. We used the command line for everything, including email, product testing and, of course, managing Solaris. The command line put the operator in control. You had a mental map of what you were doing, you were completely engaged, and if something became repetitive, you could always write a script for it. The shell was the interface, and emacs was the only tool you needed.

When GUI's first came out, we hated them on principle. They were slower than the command line, and they didn't really add any value. Plus, they weakened your skills.

Since then it's been a tossup. GUI's have gotten steadily better, but they didn't add enough value to overcome our attachment to the command line. In fact, we kinda resented them because they were used as a pretext to hire less experienced and cheaper sysadmins.

However, with the advent of vertically-integrated systems such as Oracle's Exadata and SuperCluster, the GUI may have finally come into its own. Listen to Brian Bream explain why.

Watch video interview here

Photograph of bicycle in Durango taken by Rick Ramsey in Oct 2012

- Rick

Follow me on:
Blog | Facebook | Twitter | YouTube | The Great Peruvian Novel

Thursday Sep 12, 2013

Should You Consolidate Your Servers Onto Oracle SuperCluster?

"Are you planning to consolidate a server running a business-critical application that you want to update with future releases over upcoming years, or are you trying to get rid of an old server running a legacy application that will not be updated anymore?"

This is just one of the questions Thierry asks in his article, which is a great resource for sysadmins, systems architects, and IT managers who are trying to decide whether to consolidate individual servers onto an Oracle SuperCluster. Your answer will determine whether you should put your application in native or non-native Oracle Solaris zone.

Other questions Thierry and friends ask:

  • Is my server eligible for physical-to-virtual (P2V) migration?
  • Are you planning a long-term or short-term migration?
  • How critical are performance and manageability?

Once he has helped you determine your general direction, he discusses these architectural considerations:

  • SuperCluster domains
  • Network setup
  • VLAN setup
  • Licensing considerations

Finally, he provides a thorough step-by-step instructions for the migration itself, which consists of:

  • Performing a sanity check on the source server
  • Creating a FLAR image of the source system
  • Creating a ZFS pool for the zone
  • Creating and booting the zone
  • Performance tuning

And just in case you're still not sure how it's done, he concludes with an example that shows you how to consolidate an Oracle Solaris 8 Server Running Oracle Database 10g. It's all here, give it a good read:

Technical Article: If Virtualization Is Free, It Can't Be Good, Right?

Article by Thierry Manfé, with contributions from Orgad Kimchi, Maria Frendberg, and Mike Gerdts

Best practices and hands-on instructions for using Oracle Solaris Zones to consolidate existing physical servers and their applications onto Oracle SuperCluster using the P2V migration process, including a step-by-step example of how to consolidate an Oracle Solaris 8 server running Oracle Database 10g.

Video Interview: Design and Uses of the Oracle SuperCluster

Interview with Alan Packer

Allan Packer, Lead Engineer of the Oracle SuperCluster architecture team, as explains how the design of this engineered system supports consolidation, multi-tenancy, and other objectives popular with customers.

By the way, that's a picture of an 01 Ducati 748 that I took in the Fall of 2012.

- Rick

Follow me on:
Blog | Facebook | Twitter | YouTube | The Great Peruvian Novel

Wednesday May 22, 2013

Just Add Water

When it comes to web browsing, I have little patience with amateurish sites and won’t hesitate to point out problems and flaws to the webmaster—if I can find a link to them, that is. Have you ever had this experience: you are on a web site trying to buy something and it is so slow and unresponsive that you decide to forget it: it isn’t worth the trouble or you just don’t have time? Chalk that up as one lost sale! This is—or at least should be—the web team’s nightmare.

Sure, the Web is the ultimate source of free information. But, ultimately something needs to pay the bills. Since the Web has become the universal marketplace, you would think every business would want to maximize their return by optimizing their web commerce infrastructure…

Do you remember Sea Monkeys? Just add water and soon you would have your own little amusing zoo. “Sea Monkeys” (actually dried brine shrimp) have been around since I was a kid. Advertisements were a staple of comic books. I see their purveyor, Transcience Corporation, is still in existence and taking orders!

Where is this leading? I would like to think of Oracle ATG Web Commerce as the sea monkeys of web commerce: just add Oracle’s SPARC SuperCluster. “Sure,” you are thinking to yourself, “he is paid to say that.” I can’t deny that, but my entire career in the computer industry has swirled around the dream of reusable components. We now have them: software and hardware. My enthusiasm is both justified and sincere.

A proof is in one of the best articles to cross my screen in some time: “How to Optimize Your Web Commerce Infrastructure By Tuning Oracle ATG Web Commerce Applications on SPARC SuperCluster”. This dissertation is surprisingly compact for the amount of information behind it. As the authors put it:

The objective of deploying an internet storefront or what is commonly called an e-commerce Website is to produce revenue through product offerings and subsequent purchases at the Website while at the same time “learning” about customers and their preferences. Ultimately, the goal is to make it easy for customers to research and purchase items on the site while encouraging customers to purchase related products and services. … Delivering a positive user experience also requires fast response time.

The article discussed the architecture used for testing. To get there, they had to figure out what an optimal test workload would look like and then how to simulate it. They then ask the really important question, “can this workload scale on a bigger system?”

In this study the authors developed a number of best practices, tweaks to make things run better, which they share:

  • Configuring Oracle SPARC SuperCluster with Oracle VM Server for SPARC
  • Setting Up the Oracle WebLogic Server Zone
  • Setting Up the Oracle ATG Web Commerce Zone
  • Modifying the General Purpose Domain
  • Modifying the Oracle ATG Web Commerce Environment
  • Modifying the Oracle WebLogic Server Environment
  • Modifying the Database Server

For example, because the Web server will generate a lot of concurrent processes, all accessing the database, you will want increase these in the database configuration. If you think about it, this is a meager list of tweaks for such a complex system.

Response time metrics as Oracle ATG Web Commerce workload was incrementally increased
Response time metrics as Oracle ATG Web Commerce workload was incrementally increased

This article is a great read. At the very least, it gives you an approach and methodology to testing. More importantly, it demonstrates how easy it is to create a stable and scalable solution today. Just add water.

—Kemer

Follow us on:
Blog | Facebook | Twitter | YouTube

Wednesday Aug 01, 2012

Just because I’m paranoid doesn’t mean…

KeyholeWhile I’m a passionate computer user–recognized within my family and circle of friends as a reigning (or at least old) geek–I spend at least as much time warning people to be careful as I do showing them the cool things they can do with their computers. I’m shocked at the widespread complacency over computer and network security and privacy: we should all be afraid. Very afraid. I only need remind you of recent security breaches with LinkedIn and Dropbox!

I have been in the business of publishing systems best practices for over a decade: security has been a keystone topic all those years. The good news is that high levels of security can be achieved: you just have to be smart about it. With a few exceptions, security isn’t something we actually sell, because it is dominated by relatively unglamorous concepts, processes, and practices, not extra hardware and software.

Because of this, security experts often find themselves trying to teach really fundamental (and after-the-fact, obvious) concepts. For example, in building a secure environment, you want to make sure your platform fulfills four important points:

  • It must be able to prevent or minimize the damage caused from both accidental and malicious actions. This is referred to as survivability.
  • It provides a layered set of defenses exist so that secure operations continue even after a vulnerability or the failure of a single security control. This is referred to as defense in depth.
  • It provides only those services that are absolutely necessary to the function or user. This is referred to as least privilege.
  • It is critical to be able to detect and report a breach. This is referred to as accountability.

We just posted an important article, written by three security experts I have worked with for a long time: Best Practices for Securely Deploying the SPARC SuperCluster T4–4. In building a solution that will survive a world of sophisticated cyber-criminals, it is more important than ever to pick the correct hardware and software platform. This article gives you a crash course in the things to consider, and explanation of the special features of the SPARC SuperCluster T4–4 that will make your job of creating a secure environment easier, and (most importantly) how to go about putting things together.

This is a big and important topic. Once you have digested this “Reader’s Digest” article, I’m confident that you’ll want to look at the references listed at the end. Now is the time to get smart about security.

–Kemer

About

Contributors:
Rick Ramsey
Kemer Thomson
and members of the OTN community

Search

Archives
« April 2014
SunMonTueWedThuFriSat
  
1
2
3
4
5
6
7
8
9
10
12
13
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
   
       
Today
Blogs We Like