Cisco Virtualization, the Price You Pay

We were lucky enough recently to have two comparable Cisco SAP SD benchmark results published. The results used pretty much the same hardware, the same software, the same benchmark kit. The big difference was that one was virtualized and one was not.

The performance metric for this benchmark is the number of SAP benchmark users. For the regular configuration the number of users was 6530. The virtualized version was 1000 less.(1)

That’s a considerable difference when it comes to running your business.

Compare that to the legacy results of PowerVM performance on this very same benchmark. The results of both the virtualized and non-virtualized versions are essentially the same.(2)

Power Systems servers implement a virtualization architecture with components embedded in the hardware, firmware and operating system software. The capabilities of this integrated virtualization architecture are thus significantly different and in many areas more advanced.

Without paying the performance price.

************************************************

(1) SAP SD 2-tier Cisco UCS B200 M3, 2 processors / 16 cores / 32 threads, Intel Xeon Processor E5-2690, 2.90 GHz, 64 KB L1 cache and 256 KB L2 cache per core, 20 MB L3 cache per processor, 256 GB main memory; SAP SD benchmark users: 6,530 , Average dialog response time: 0.98 seconds, Throughput: Fully processed order line items/hour: 713,670, Dialog steps/hour: 2,141,000, SAPS: 35,680. Average database request time (dialog/update): 0.015 sec / 0.036 sec, CPU utilization: 99%, Red Hat Enterprise Linux 6.3, Sybase ASE 15.7, SAP enhancement package 5 for SAP ERP 6.0, Certification #2013001.
SAP SD 2-tier Cisco UCS B200 M3, 2 processors / 16 cores / 32 threads, Intel Xeon Processor E5-2690, 2.90 GHz, 64 KB L1 cache and 256 KB L2 cache per core, 20 MB L3 cache per processor, 256 GB main memory; 1 virtual machine (VM) using 32 virtual CPUs. CPU utilization of VM1 (DB/Dia/Upd/Msg/Enq): 97%, Number of SAP SD benchmark users: 5,530; Average dialog response time:0.96 seconds;Throughput: Fully processed order line items/hour: 605,330; Dialog steps/hour:1,816,000;SAPS:30,270; Average database request time (dialog/update):0.021 sec / 0.045 sec; CPU utilization of central server:97%; Operating system, central server:Red Hat Enterprise Linux 6.4 on KVM; RDBMS: Sybase ASE 15.7; SAP enhancement package 5 for SAP ERP 6.0; Certification #2013007.

(2) SAP SD 2-tier IBM Power 570, 2 processors /4 cores / 8 threads, POWER6, 4.7 GHz,Number of SAP SD benchmark users: 2035, users/core = 508.75; Average dialog response time: 1.99 seconds;Throughput:Fully Processed Order Line items/hour: 203,670;Dialog steps/hour: 611,000;SAPS: 10,180;Average DB request time (dia/upd): 0.011 sec / 0.015 sec;CPU utilization of central server: 99%;Operating System central server: AIX 5L Version 5.3;RDBMS: Oracle 10g;SAP ECC Release: 6.0;Certification #2007037.
SAP SD 2-tier IBM Power 570, 2 processors /4 cores / 8 threads using 2 virtual cpus, POWER6, 4.7GHz; Number of SAP SD benchmark users: 1020, users/core=510;Average dialog response time: 1.99 seconds;Throughput:Fully Processed Order Line items/hour: 102,000;Dialog steps/hour: 306,000;SAPS: 5,100;Average DB request time (dia/upd): 0.005 sec / 0.009 sec;CPU utilization of central server: 50%;CPU utilization inside virtual machine: 99%;Operating System central server: AIX 6.1 on IBM Power VM (using 2 virtual CPUs);RDBMS: DB2 9.5;SAP ECC Release: 6.0; Certification #2008080

http://www.sap.com. Results current as of 3/20/13.

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

SAP and all SAP logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries. Other names may be trademarks of their respective owners.

technorati tags: IBM, systems, performance,SD,Cisco,UCS,B200,virtualization,POWER,PowerVM,storage,DB2,benchmark,benchmarking

IBM, Easy Being Green

Sometimes you just want to dance to what’s trendy. Like Gangnam Style in July, or the Harlem Shake.

Many organizations are like that as well. Let’s offer work at home (now let’s not). Free food in the cafeteria. Let’s be green now.

In the latest news, IBM was just presented with a 2013 Climate Leadership Award in Washington, D.C. by the EPA, the Association of Climate Change Officers, the Center for Climate and Energy Solutions, and The Climate Registry.

This is not a one-off. A let’s work on it this year. A Beanie Baby, Cabbage Patch anomaly.

IBM has been doing this for over 40 years.

Through initiatives such as green data centers, smarter buildings, virtualization and consolidation of systems, cooling technologies.

And a good thing too — as yesterday scientists reported that global temperatures are the highest in 4000 years. Yes, that’s 4000 years. And over the coming decades they are likely to surpass levels not seen on this planet since before the last ice age.

************************************************

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: IBM, systems, performance,green,EPA,energy,efficiency,virtualization,climate,consolidation,

storage,smarter,benchmark,benchmarking

Oracle, Blowin’ in the Wind

Oracle announced this morning the general availability of the Oracle Database Appliance X3-2. It now supports virtualization.

Hmmm, not exactly groundbreaking dare I say.

Along with this Oracle announcement were many claims of bigger and better — than a previous version.
With no real data and no real comparisons. And all I could think of was a Saturday morning a couple of months ago.

You see, I was in an industrial supply store this one Saturday morning (the things we do for love). Looking at, of all things, leaf blowers.

• Some models of these new leaf blowers claimed to blow harder that others. Surely harder than the old broken one at home. Which never really blew all that well. In fact, I probably ended up using my beloved rake more than that big hunk of plastic.
• These new leaf blowers claimed to be bigger and better. That meant that they actually weighed a ton more and had to be converted from a hand held model to a huge backpack. Ghostbuster time. And you needed earplugs.
• They let you try it out in the store. But instead of heavy wet leaves, they had you blow strips of paper.
• The advertised pricing did not include the expensive oil that you needed for it — without that specific oil, the machine would die.

So it comes down to the fact that claims are similar all over. But does the appliance finish the job fast, reliably, and at the right total price — so that you can get to what you really want to be doing.

************************************************

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: IBM, systems, performance,Oracle,appliance,database,X3-2,virtualization,VM,SAS,storage,

WebLogic,benchmark,benchmarking

Oracle, I Knew You Were Trouble When You Walked In (to SPECjbb2013)

I’m always on the lookout for anything that helps me eat well. My newest interest these days is a type of rye crisp cracker that looks like cardboard, tastes a bit better, but has absolutely no fat. (They are really good if you put a huge slice of brie on them.)

I also love reading about nutrition. I love Bloomberg’s ban on big soda. I love the concept of using color to focus on a well-rounded and healthy diet. Does lime jello count in the green group?

Pomegranate juice, which has many claims on being good for you, was one of those things. Deep red, and also great for an amazing sauce reduction for chicken. Well, I recently read that the Federal Trade Commission just barred claims of pomegranate juice helping heart disease and other ailments until truly proven — based on deceptive advertising claims.

Which reminded me of something else. Oracle recently claimed a “world record” in the SPECjbb2013 Java benchmark. What you need to know is that this benchmark does not have 100 other results to compare to. It doesn’t have 20. It doesn’t even have 10. This “world record result” is based on just two other published results. Both by Intel. With older JDK. And fewer and older processors.

Like when your black lab starts to sit on your friend’s chihuahua. Or a claim that your juice cures cancer.

************************************************

SPEC, SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPECjEnterprise, SPECjvm, SPECvirt, SPECompM, SPECompL, SPECsfs, SPECpower, SPEC MPI and SPECpower_ssj are trademarks of the Standard Performance Evaluation Corporation (SPEC).

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: IBM, systems, performance,Solaris,JDK,SPEC,SPECjbb,SPECjbb2013,Intel,E7,Sun,X2-4,

benchmark,benchmarking

T4 MIA

I was recently looking to do a comparison of a couple of systems and figured the easiest data to look at would be SPEC CPU data. I mean it’s almost a given that even though these integer and floating point benchmarks can be, should I say, pretty boring (yawn), at least you’ll find what you are looking for because everyone publishes pretty much everything on these.

To my surprise I realized that I did not see the one system I really needed to look at – a SPARC T4 result. And then I started to laugh to myself (it was pretty late).

Of course. T4 is the one where you can find results for some really strange benchmarks but none of the regular ones where you can actually compare anything.

And then I remembered something else. I went out to my blog just to confirm it and there it was staring at me in the face. The #1 all time most popular search term that gets people to read my blog is just that — “sparc t4 benchmark.” Followed closely by “sparc t4.” “sparc t4 specint.” “sparc t4 benchmarks.” “specint t4.” “t4 specint.” And “sparc t4 performance.”

Pretty sad that for this type of basic data you can not find it on the SPEC site. You can not find it on the Gartner site. You can not even find it on the Oracle site. For this type of basic performance data you need to go to an IBM site to find out that this data just doesn’t even exist.

************************************************

SPEC, SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPECjEnterprise, SPECjvm, SPECvirt, SPECompM, SPECompL, SPECsfs, SPECpower, SPEC MPI and SPECpower_ssj are trademarks of the Standard Performance Evaluation Corporation (SPEC).

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: IBM, systems, performance,SPARC,T4,SPEC,SPECcpu,CPU2006,SPECfp,POWER7+,SPECint,Power,benchmark,benchmarking

Life in the Fast Lane

I recently wrote about a college visit — Big Data, Performance, and Coconut Smoothies might as well be titled Big Data at the Big Ten. And the ability to collect, measure, and analyze this massive flood of data for meaningful insights requires important non-functional IT requirements like reliability, availability, security, and of course performance.

The IBM announcement of new products and offerings today leverages cloud to improve efficiency, focuses on data to deliver more actionable insight, and secures this critical data to protect and reduce risk. It’s Cloud Ready, Data Ready, Security Ready. And, of course, Performance Ready with:

• New #1 Power Systems benchmarks for the new POWER7+ systems across a wide portfolio of applications including SAP, Java, and Technical Computing. Read about them here.
• All the benchmarks and rPerf data in the performance report here.
• Proof points like this one: The new 48-core IBM Power 760 with DB2 10 achieved the best 48-core two-tier SAP SD standard application benchmark result, surpassing the 80-core HP DL980. (1)
• Performance details on the new POWER7+ systems including Power 730, Power 740, Power 750, Power 760, PowerLinux 7R2.
• Leadership Storage performance including performance enhancements to IBM XIV. New caching algorithms increase performance compared to previous models up to 4.5 times for random and 5 times for sequential database workloads.

One thing that struck me on this college visit (besides how young everyone looks) is how college truly is an exciting new beginning. Everything is shining, brand spanking new. But this doesn’t only have to happen in college. There are frequently new beginnings throughout life.

Years ago, when social media was a relatively new concept for many of us, my manager at the time (who was always a source of exciting, new, and crazy ideas) “suggested” that I consider writing something called a blog. A downright scary idea, I implemented it anyway.

This week, this esteemed manager, colleague, mentor, associate is retiring after a very successful career of exciting ideas. To start a new beginning in the fast lane. And I will miss his inspiration. And his crazy ideas. Like a blog.

************************************************

(1)The 48-core IBM Power 760 (3.41 GHz) achieved the best 48-core two-tier SAP SD standard application benchmark result running SAP enhancement package 5 for the SAP ERP 6.0 application; 8 processors / 48 cores / 192 threads, POWER7+, 1024 GB memory, 25,488 SD benchmark users, running AIX® 7.1 and DB2® 10, dialog resp.: 0.99s, line items/hour: 2,784,330, Dialog steps/hour: 8,353,000 SAPS: 139,220, DB time (dialog/ update): .009s/.015s, CPU utilization: 99%, Certification #2013004. vs. HP DL980 G7, Xeon, SQL Server 2008, 25,160 users, 8 processors/80 cores/160 threads, SAP enhancement package 4 for SAP ERP 6.0, Certification # 2011021. http://www.sap.com. Results current as of 2/5/13.

SAP and all SAP logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries. Other names may be trademarks of their respective owners.

SPEC, SPECint, SPECfp, SPECjbb, SPECweb, SPECjAppServer, SPECjEnterprise, SPECjvm, SPECvirt, SPECompM, SPECompL, SPECsfs, SPECpower, SPEC MPI and SPECpower_ssj are trademarks of the Standard Performance Evaluation Corporation (SPEC).

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: IBM, systems, performance,SAP,SD,SPEC,SPECjbb,Linpack,SPECfp,POWER7+,

SPECint,760,DB2,740,730,750,Power,benchmark,storage,benchmarking

Thinking About Tape

I was doing errands this weekend when I turned on my car radio. American Top 40. A special on the top 40 from all of the 70′s.

So there I was singing to Simon and Garfunkel – It’s not corny because my mom went to school with Art. Knocking with Tony Orlando and Dawn — those yellow ribbons really became historical icons. Even getting into Debby Boone’s “You Light Up My Life” — so sweet it made your teeth ache — which somehow made it to #1.

And all I could think of at that moment was tape. Eight-track tapes. And my big black box with the fake leather, cool snap, and plastic handle that I used to lovingly transport my eight-tracks every time I needed to take them somewhere.

Later in life, I also carried tapes. Big heavy tapes, holding the lifeblood of a company, that needed to get from an organization’s operations center to Wall Street headquarters. Loaded down with them, all I could think of at that moment was tape too.

I got to thinking about tape again last week when I read that Oracle announced new midrange tape drives available for tape libraries. All sorts of claims were made. But when the talk is about efficiency, I really want to see some numbers and some data. But I saw nothing, nada — I was truly disappointed.

Whereas IBM has published a detailed white paper on LTO-6 vs. LTO-5 performance. This white paper describes the performance of the IBM® System Storage® LTO Ultrium 6 tape drive (LTO-6 tape drive) working on open systems environments. With a ton of real data, real tests, and benchmarks. And lots of great graphs highlighting all the excellent performance.

Reading it, once again, all I could think of at that moment was tape.

************************************************

The postings on this site solely reflect the personal views of the author and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.

technorati tags: IBM, systems, performance,LTO,storage,StorageTek,Oracle,tape,benchmark,LTO-6,benchmarking