Posts by PowerMan@thinksis 88 publicly visible posts • joined 31 Jul 2013
@MadMike seems to be nearly identical to the post of "kebabbert" at EnterpriseTech's article. Read his comment and my response. He is spewing nearly the same FUD there. He is a sharpshooter who shoots his mouth off with talking points just like Phil Dunn of Oracle does. http://www.enterprisetech.com/2014/08/13/oracle-cranks-cores-32-sparc-m7-chip/#comment-229928
Um, not true. Your comment is FUD. Each socket in the current Power8 Scale-out server is package with up to 2 x 6 core chip modules. IBM has done this on Power5, Power6 and now with Power8 servers so nothing new. Moreover, whats wrong with it if it performs? Your comment about "Performance not good" is off base. What do you base it on? I would point you to the following benchmarks where you see a 24 core S824 match a 4 socket 60 core Ivy Bridge EX v2 server in SAPS & Users. Outperform in SPECint, SPECfp, SPECjbb and more. Those are just benchmarks, my customers are seeing the performance and more.
"So they went to giant off chip(s) shared L4 cache." Really? Adding technology and innovating is now a gimmick? By this explanation having L3....L2 and even L1 cache are all gimmicks. Just main memory and cpu for you. Come on, are you a bit jaded by your SPARC love? It's ok to say your Ford SPARC is the best ever but don't lie about my Chevy Power :) Power8's L1 (D+I) are 2X greater than x86 and 4X+2X over SPARC T5. L2 cache is 2X over x86 and 4X over SPARC T5. L3 is 2.5X over x86 and 12X over SPARC T5. Neither x86 or SPARC have L4 while Power8 has 128 MB per socket. This gives Power an advantage to get data closer to the core so it may fit entirely in a lower cache line.
Using your explanation of the Power8 roadmap, they began shipping Power8 in June '14. If they start shipping Power8+ (reasonable to expect IBM to stick with the entry level roll-out as they did with Power8) in 18 months that would put them the Nov / Dec '15 timeframe. Oracle having a 2015 rollout could be January or December. I think the point is valid that they are behind and will be further behind. With regard to Intel - not sure what planet you are on but I see Intel heavily focused on 4 socket servers. Ivy Bridge EP (E5) v2 for the 2 sockets and EX (E7) v2 for 4 sockets and above. I am not fully briefed on Haswell and Broadwell but it's reasonable to expect one of them will deliver a 4 socket solution. It's possible Intel is figuring out enterprise customers don't like rapid change in chipsets that run heavy duty workloads and would rather have a reliable chipset over the latest and greatest every chip release. What you cite is what Intel is battling in general - trying to go after the enterprise space while continuing to own the 2 socket space and defend against ARM in both the 2 socket but also the mobile & portable space. Google using Power, Apple considering a change plus their continued growth with the iPhone / iPad puts pressure on every chip manufacturer.
The most dense CPU ever? An easy search on google will show you this is not the most dense CPU ever created - Intel, Nvidia, Azul and more have far higher core density. Oracle isn't doing things Sun was not capable of....moreover, I would argue they are pulling old plays out of the Sun playbook in a desperate attempt to remain relevant. One example is the similarity the M7 has to the Rock processor cancelled in 2009/2010. Rock was a 16 core CPU design made up of 4 clusters of 4 cores each. Very similar to the M7 which has 8 clusters of 4 cores each - coincidence? Hmmm!
Any idea what the OS and server platform that was infected with the Malware was? This kind of data is crucial for readers to understand the most prolific and lowest cost solution may expose them to criminal and civil liabilities. If they take on the responsibility of handling customer data they need to go with a reliable and *secure* platform from the start.
Unfortunately IDC might as well be tracking snipe if we are to believe this. IBM's 25 year old iSeries platform is the iconic integrated platform. As a seller of the iSeries I am confident the numbers they show for IBM are far from accurate and not shown in their data or would be far greater than Oracle's "Integrated Platform". They don't disclose what they do represent in the article or at IDC's website or in their press release at http://www.idc.com/getdoc.jsp?containerId=prUS24963914. My guess is it represents IBM's PureApplication and maybe PureData only. Also, I've yet to give away a iSeries to a customer but have several customers where Oracle has given them an "Exa*" product to seed it. Can't help but wonder how much of their numbers represent these free and cheap x86 white boxes loaded with software that is free when delivered and doesn't cost the customer until they start paying 22% maintenance on all of those x86 cores on year 1....but at least it is only with a core factor of .5!
Also, while I'm at it, VCE nor FlexPod are not converged technologies by definition of "convergence". Just because they plug EMC storage into a Cisco switch which connects to a UCS chassis then install VMware on a UCS node is NO different than taking a V+C+E solution and doing the same thing. Vendors and partners have been integrating technologies for years so VCE is doing nothing novel. The only technology stack that was truly "converged" was IBM's PureSystems with its Flex chassis, integrated management, integrated server nodes, integrated storage nodes, integrated SAN & Ethernet switches that all communicate inside the Flex chassis. The next technology that is more converged than VCE or FlexPod is IBM's Power servers. With their integrated hypervisor and ability to bring Ethernet & SAN into the server then share among hundreds of VM's to multiple OSes, with internal firewalls, encryption, management, etc. IDC demonstrates they do not understand the available technology and instead buys into the hype of these aggressive marketing machines.
That explains why I couldn't buy a USB drive at my local Fry's this past weekend - HP is buying up all of the cheap Flash using the consumer grade MLC rather than the Enterprise grade eMLC or SLC. I'm being sarcastic for those who might get bent out of shape by my comparison of USB drives with Flash vs the Flash used with SSD drives - yes, both use Flash technology but in different form factors. Heck, HP even knows better than the SSD vendor and is having them cut back on over-provisioning capacity - Whatever it takes to get the headline I guess. You know how there is a uproar in the grocery store over food makers using words like Lite, Lo-Calorie, Fat Free, Low Fat, Orgranic, etc? We need a oversight organization to straighten out all of the "marketing" liberty these vendors are taking with "Flash". All "Flash" is not created equal and I think this HP 3PAR may not be all that it claims to be.
Now we see why performance/watt matters - trade-off of compute density vs the systems ability to not fail. Pretty straight forward. It continues to highlight the general weakness of the x86 chipset and lack of technology innovation with x86 vendors in general - HP in this case. IBM had their Power7 775 supercomputer in 2011 which had 12 x 256 core 775 servers in a rack - that is 3072 cores - the solution itself can scale up to 512K cores. IBM seemingly left the market for these very expensive demonstrations of chest thumping - thank goodness. I'd rather see them invest in training and marketing efforts to educate more IT shops and businesses on better options to x86. But, it makes me wonder why HP with all of their own financial troubles would want to jump into these shark infested waters.
Of course, makes complete sense that VMware offer a hardware solution as only they would know best how to exploit their products. Just like Oracle with their Exa* products - they know best how to optimize their software stack......and VCE with their vBlock as only they would know how to integrate the three tiers of the infrastructure to optimize the stack. Hopefully you picked up on the sarcasm. VMware should do what it does best which is software and from my perspective with the virtualization stack. If they have extra time they could tighten up their security. Oracle says they optimize the stack with Exa* only to be 75% marketing and 25% technology. White box x86 servers packed with software licenses (and lots of cost) and numerous gotcha's - extra storage and other features to address the solutions many inherent weaknesses. VCE, nothing more than a infrastructure integrator that upcharges V+C+E products where they tout flexibility but in reality are rigid - once implemented you make changes according to their schedule nor your business. Can't wait for yet another hardware player that has the answer to optimizing the stack!
Hey anonymous coward, if you only talked about how great your "Intel kit" was then you couldn't be blamed for simply liking what you like. But then you go and start spewing FUD and lies about IBM Power like you are an authority.
If Intel is so "open", why did Google and 27+ other major companies join IBM in the OpenPower Foundation? That is where innovation is taking place and why? Why, because Intel is actually shutting out partners and stifling innovation as they try to expand and own it all - kind of like Oracle is doing by buying hardware, storage, applications, middleware, OS, support, backup, etc. Unlike Oracle which is in the enterprise, Intel extends from the consumer to commercial.
OpenPower with Google, Tyan, Mellanox, Nvidia, Samsung, Xilinx, Micron, Hitachi, Ubuntu and more are working on some crazy products that will deliver features and capabilities that stagnant x86 won't know what hit them. If you want to see some innovation go read up on them.
Do you even know how much a Power server cost? You write like you are just "put off" that there could be an option to your much beloved x86 platform, let alone it be competitive. Your first mistake is pricing a core for core Power server to a x86 server. I don't need a 2 socket 16 core Power8 server to do the same work as a 2 socket 16 core Intel server. But, you (x86 fanboi's) will try to make that comparison all day. Let me point you to the SAP S&D 2 tier benchmark comparing the 2 socket 24 core Power8 S824 server to the 4 socket 60 core NEC server that were both posted at the end of April. The 4 socket IB E7 v2 which you say the "latest Intel kit far exceeds anything that IBM is offering" generated 114,700 SAPS with 20,800 Users vs the 2 socket 24 core Power8 S824 with 115,870 SAPS and 21,212 Users. There are a couple of 4 socket E7 v2 that go up to 135K but you can't dismiss the incredible performance of the "latest IBM kit (actually) far exceeds anything that Intel is offering". The P8 has 4828 SAPS per core compared to just 1912 SAPS for the E7 v2 or should I call it the "latest Intel kit"? If that doesn't work for you how about the SPECint for the 24 core Power8 at 1750 vs 1020 for a 24 core E5 v2 - yes, now I am comparing 2 socket to 2 socket and core for core. SPECfp is 1370 for P8 and 734 for E5 v2 as well. Would you like to hear the SPECEnterprise2010 result? How about that it is 939 EjOPS per P8 core compared to 469 for the E5 v2 core. I could go on and dive deeper into how PowerVM delivers QoS for it's VM's, allows for dynamic changes of cpu, memory and I/O, concurrent updates to firmware and other maintenance, etc, etc. Besides just performance there are the benefits of the integrated stack. Power still uses partners to deliver products like memory, adapters and disk drives - they just ensure they do so at tight tolerances and specification. Further, they test them to work with the hypervisor, OS(es), HA products, filesystems, storage, etc so the user experience is positive and the results are exceptional - doesn't mean they are perfect but anybody who argues a Power server is not reliable has NO experience working with one - I don't know you but can say that without hesitation. You can hate the platform, dislike IBM and love your ugly baby but you don't get to make up the facts.
Your utilization claims are suspect for sure - some shops over 80%....hmmm, I call BS. If they are x86 I doubt they are virtualized. I think you are trying to be provocative again. I'll do some research for recent studies by all vendors to see what they say is becoming the norm - I hardly believe 80% is. When you respond with more FUD, please provide some links to a vendor or even well known industry bloggers who have discussed this.
I already refuted your performance claims above showing you just make stuff up to suit your argument. IBM has Power8 servers with price parity to x86 servers - not all of them but definitely in the entry space. When sized for the workload, including the virtualization, OS and 3 years of support they are virtually equal. Just like x86 vendors and the unaware want to do by comparing a 16 core Power to a 16 core x86 server, you also look at the TCA or acquisition price as if that is the relevant factor. It is 1 of about 20 components that make up the solution cost. There is the data center power, rack & cooling costs. LAN & SAN cabling & port costs, OS & virtualization license & support costs, security products like anti-virus, HA or other clustering products like Oracle RAC at $23,000 per core + 22% maintenance per year. Application and middleware costs. You don't want to hear this and won't believe it as you probably do not think the rich pay enough taxes, that America is the cause of the world's problems or that x86 is just as reliable or secure as any RISC/Unix or mainframe platform. But, a properly sized Power solution running Oracle for example against an appropriately sized Intel solution (pick your vendor) with Linux will cost 4X of the Power solution. The server cost is a nit as the largest portion of the TCO is software and we haven't even brought up the increased number of FTE's required to support the proliferation of heat producing toasters in the data center that require non-stop attention to replace disk drives, memory, power supplies - having VM's all over the place as they vMotion themselves as soon as the server see's a increase in utilization - oh, that's right - not your servers because they run up to 80%. Well, every other shop but yours scooter.
Interesting when people who comment with such authority yet hide behind an anonymous coward id. You've earned your title today because you are just throwing stones to agitate. I'll briefly lay out the facts to refute your cowardly and false statements. 1) Multiple factors are impacting every vendor which are back to back recession's. In Sun's case it goes back to 9/11 which they could never recover from. 2) Management or business decisions - Sun and HP are recent examples. Jonathan Schwartz took a good idea at the wrong company at the wrong time. HP has had a succession of chaos since arguably Carly. The Leo reign was just horrible. If they didn't have full chaos he sure created it. 3) Financial results are disproportional to Technology improvements. As x86 improved - thanks to Itanium input and AMD embarrassing Intel, Linux made it good enough. TCA or Total Cost of Acquisition has become the biggest smoke and mirror game going. Virtualization is the shiny object making it worse. Big Iron vendors like Sun & HP disinvested leaving IBM to carry the RISC mantra. IBM has continued to invest in R&D with Power which has seen increased performance per core (could show you the last 5 generations as a proof point going back to 2001 if you like) and more RAS & virtualization features. Just for fun let's look at Itanium - stuck at 1 & 2 cores running at 1.5 & 1.6 GHz Itanium for years before they finally made the leap to 4 cores. Have they ever officially said Itanium was dead? Word is they will tell customers under NDA but those features are being moved into x86 and to hang on. Oracle - let's see - they have had Intel, AMD, SPARC, SPARC T series, SPARC64 and now it seems they have settled on the latter two although all of the news reports say Oracle reps loathe selling a Fujitsu box regardless if it is best for the customer. Is that because they fear loss of account control or less of a commission? Their T series languished for years at 1, 1.2 and 1.4 GHz while they had 4, 8, 16 cores per socket but they told the world they were the best at everything. When they announced the T5 they called it the "Worlds Fastest Processor" - hmm, Kevin Closson has a nice article shredding this at http://kevinclosson.wordpress.com/2013/04/09/my-first-words-on-oracles-sparc-t5-processor-the-worlds-fastest-microprocessor/. Don't forget that Power8 will now run Linux in little endian mode as well as big endian mode. It outperforms x86, it has top to bottom security. It's virtualization has zero CVE compared to x86 alternatives. AIX is feature rich and dynamic allowing for many concurrent features not available on Linux. Linux on Power is very good but AIX is best with IBM i being even better. By the way, this doesn't mean a Solaris on SPARC server doesn't run well, isn't reliable...same for HP-UX on Itanium - I respect both technologies. But they don't even compare to Power and x86 with Windows (seriously) with Linux doesn't even get an honorable mention.
Here is what it comes down to. Power8 is relevant and competitive. It has more RAS, virtualization, security and out performs EVERY platform out there core for core and socket for socket. x86 and SPARC sellers will state .5 vs 1.0 for Oracle (for example) and say Power cost 2X over x86 and SPARC. That is BS and deceptive. Power is not only powerful but efficient. You can allocate just the cores required to a VM and license those. That number will be anywhere from 3-4X fewer than SPARC. For x86, it doesn't matter - virtualization or not because you license the entire server. There are plenty of other articles out there that discuss the truth about this if you look hard enough. Read this for details http://bartsjerps.wordpress.com/2014/05/12/oracle-vmware-sub-server-partitioning/
Unless those PC's are desktops and laptops your business will soon look like HP's. Data Centers are expanding and bursting at the seams due to the horizontal scale-out and low utilization of x86 servers. Despite virtualization efforts there are claims of improvements but I don't how how anybody can claim things have improved with a average utilization of 30%. Data Centers are full of heat producing devices all with redundant power, multiple Ethernet & SAN cables running at 30% - what would your performance report say if you worked at 30% efficiency? How profitable would your business be - staffed at 100% but producing at 30%.
You should take a look at IBM's new RISC server called Power8. It runs Linux and AIX. It can almost do with their 2 socket 24 core server what it took 60 cores of HP's 4 socket DL580 with the latest IB E7 v2 processor. If you are running software that is price by the core not to mention the data center costs for power, cooling, rack space, switch ports plus all of the software, maintenance and FTE's to support all of this....well, I have seen consolidations of 10:1, 20:1 even 50:1 - that is servers by the way. The article isn't about this but your write-up in support of the company just begged for somebody to step in and comment!
You would be right if this was Power6 days. It is not. With Power7 and now Power8 there are price parity options for TCA and when talking Oracle they dominate when it comes to TCO. Want to see how Oracle on Power is less than x86 ….. Read on! I'll pick a random x86 vendor that will let me get pricing from their website . I use list prices because they are consistent in all comparisons. Discounts can be applied but it doesn’t change the ratio of savings, just how much.
.
HP DL380 Gen8 - qty 1
2 x 6 co @ 2.4 GHz E5-2440 totaling 12 cores
128 GB Ram
vSphere Enterprise Plus
Assume Linux – pick your distro of choice
3 year support
No internal HDD - assuming USB boot
2 x dual port 10 GbE
2 x dual port Fibre
All power cores, rail kit, misc
$25,183 each server List price
.
.
Oracle Licensing cost
Enterprise Edition - $47,500 per core
EE maintenance @ 22% per year - $10,450 / co
RAC - $23,000 per core
RAC maintenance @ 22% per year - $5,060 / co
.
.
IBM’s latest Power8 server
S824 Power8 server - qty 1
(this is the 2 socket model although I just selected it with 1 socket)
.
8 x 4.15 Ghz Power8 cores
256 GB Ram (need more Ram because I’m planning to host more VM’s)
DVD
Split backplane
4 x SSD (building the way I would built it and not just to lower the price which I could do by using HDD)
2 x dual port 10 Gbe adapters
2 x dual port Fibre adapters
AIX v7.1
PowerVM Enterprise Edition
3 year 24 x 7 maintenance
$79,807 server list price
.
.
Now the math and the comparisons!
.
Server: HP DL380
Cost: $25,183
qty of servers: 2
Server cost: $50,366
.
# of cores in solution: 24 (2 x 12 co servers)
Oracle Licensing Factor: .5
# of cores needed for Oracle (actual): 5 (nice to know but doesn’t matter for licensing)
Total Oracle Licenses required 12 because (24 * .5 = 12)
.
Oracle EE Lic cost: $570,000 (12 * $47,500)
Oracle maint cost (3 yr): $376,200 (12 * $10,450 * 3)
Oracle RAC Lic cost: $276,000 (12 * $23,000)
Oracle maint cost (3 yr): $60,720 (12 * $5,060 * 3)
Total x86 server + Oracle cost over 3 years: $1,333,286 (Add it all up)
.
.
Server: S824
Cost: $79.807
qty of servers: 1
Server cost: $79,807
# of cores in solution: 8
Oracle Licensing Factor: 1.0
# of cores needed for Oracle (actual): 3 (it does matter here)
Total Oracle Licenses: 3 (either in a dedicated or SPP with proper boundaries)
Oracle EE Lic cost: $142,500 (3 * $47,500)
Oracle maint cost (3 yr): $94,050 (3 * $10,450 * 3)
Oracle RAC Lic cost: $ Not required
Oracle maint cost (3 yr): NA
Total Power server + Oracle cost over 3 years: $316,357
.
.
The 3 year total cost of ownership for the x86 solution shown is $1 Million dollars more than a Power8 solution.
.
Somebody may question or say it isn’t fair or that it is convenient of me to just use 1 Power8 server whereas I am comparing it to 2 x HP x86 servers. Just in case, here are those numbers. Don’t want somebody to not like me accuse me of making things up ☺.
.
Using 2 x Power8 servers instead of 1 to compare:
.
Server: S824
Cost: $79.807
qty of servers: 2
Server cost: $159,614
# of cores in solution: 16 (2 * 8 cores)
Oracle Licensing Factor: 1.0
# of cores needed for Oracle (actual): 6 (because each server has 3 co in a Active / Active cluster)
Total Oracle Licenses: 6
Oracle EE Lic cost: $285,000 (same math as above)
Oracle maint cost (3 yr): $188,100
Oracle RAC Lic cost: $138,000
Oracle maint cost (3 yr): $91,080
Total Power server + Oracle cost over 3 years: $861,794
.
.
For those who want everything equal the Power solution is still $470K less than the x86 and everything else I have said remains true.
Best way to tune Oracle is to run it on a Power8 server with IBM FlashSystem 840. Reduce Oracle licenses to 1/4th of any other platform at higher performance than x86 and SPARC eliminating I/O latency with Flash. Further, to the other persons point, Oracle nickel and dimes customers on every feature. DB2 v10.5 with BLU includes many of these features like compression, replication, HA, tuning, their RAC equivalent and more for less money. And, the software license includes the first year of software maintenance unlike Oracle where you pay 22% for year 1 plus the software license cost. If SQL Server also offers similar features and pricing that is great - gives customers more choices. For SMB shops maybe go the SQL Server route. For customers with demanding, enterprise needs there is nothing stronger than a Power8 server.
Now the math and the comparisons!
.
Server: HP DL380
Cost: $25,183
qty of servers: 2
Server cost: $50,366
# of cores in solution: 24 (2 x 12 co servers)
Oracle Licensing Factor: .5
# of cores needed for Oracle (actual): 5 (nice to know but doesn’t matter for licensing)
Total Oracle Licenses required 12 because (24 * .5 = 12)
.
.
Oracle EE Lic cost: $570,000 (12 * $47,500)
Oracle maint cost (3 yr): $376,200 (12 * $10,450 * 3)
Oracle RAC Lic cost: $276,000 (12 * $23,000)
Oracle maint cost (3 yr): $60,720 (12 * $5,060 * 3)
Total x86 server + Oracle cost over 3 years: $1,333,286 (Add it all up)
.
.
Server: S824
Cost: $79.807
qty of servers: 1
Server cost: $79,807
# of cores in solution: 8
Oracle Licensing Factor: 1.0
# of cores needed for Oracle (actual): 3 (it does matter here)
Total Oracle Licenses: 3 (either in a dedicated or SPP with proper boundaries)
Oracle EE Lic cost: $142,500 (3 * $47,500)
Oracle maint cost (3 yr): $94,050 (3 * $10,450 * 3)
Oracle RAC Lic cost: $ Not required
Oracle maint cost (3 yr): NA
Total Power server + Oracle cost over 3 years: $316,357
.
.
The 3 year total cost of ownership for the x86 solution shown is $1 Million dollars more than a Power8 solution. The Power solution is very typical for what we might see or use with customers. We would also consolidate the app servers and other workloads onto the Power server whereas customers typically would put the App servers on separate servers - which means even more cost.
.
Somebody may question or say it isn’t fair or that it is convenient of me to just use 1 Power8 server whereas I am comparing it to 2 x HP x86 servers. Just in case, here are those numbers. Don’t want somebody to not like me accuse me of making things up ☺.
.
Using 2 x Power8 servers instead of 1 to compare:
.
Server: S824
Cost: $79.807
qty of servers: 2
Server cost: $159,614
# of cores in solution: 16 (2 * 8 cores)
Oracle Licensing Factor: 1.0
# of cores needed for Oracle (actual): 6 (because each server has 3 co in a Active / Active cluster)
Total Oracle Licenses: 6
Oracle EE Lic cost: $285,000 (same math as above)
Oracle maint cost (3 yr): $188,100
Oracle RAC Lic cost: $138,000
Oracle maint cost (3 yr): $91,080
Total Power server + Oracle cost over 3 years: $861,794
.
.
For those who want everything equal the Power solution is still $470K less than the x86 and everything else I have said remains true. If the Oracle workload grows and needs more resources, the Power server can dynamically add a single (or fraction of) core at a time and any increment of memory to the VM. You just add the appropriate Oracle licensing. Likewise, if the workload were to decrease you could dynamically remove cores and memory as well and even redeploy Oracle licenses to other workloads or other servers - a license at a time. It is all about flexibility.
.
Heaven forbid you had a x86 Oracle RAC environment Plus VMware with more servers than are configured with RAC where you would license Oracle at $70,500 per core across all of the cores in that vMotion cluster farm. In the example I have shown above though, it will cost significantly more to run Oracle on x86 with or without VMware vs running it on Power. Cheers!
IBM’s Power servers are the only platform to control Oracle licensing cost. With x86 the performance seems to go down with each release while the overall socket performance goes up because the total number of cores increase. With Power server, each successive generation of server delivers more performance per core so for a static workload the overall licensing requirement would go down.
.
For a typical x86 solution to host an Oracle Enterprise Edition database, customers would not deploy Oracle on x86 in production without some form of increased availability like Oracle RAC. This adds $23k/co to the existing $47.5k/co price for the EE DB. I use list because they are consistent in all comparisons. Discounts can be applied but it doesn’t change the ratio of savings, just how much.
.
With the Power server, additional clustering is not required to address or overcome the inherent deficiencies in the server platform like it is with x86. It is often included though for business reasons to ensure all reasonable measures have been taken to avoid an outage. If increased availability is required, customers may choose to go with a traditional cluster product like VCS (I still call it this) or something like PowerHA, which I prefer as it is less expensive and more robust. Customers may also choose to use Oracle’s RAC.
.
Want to see how Oracle on Power is less than x86 ….. Read on! I'll pick a random x86 vendor that will let me get pricing from their website
.
HP DL380 Gen8 - qty 1
2 x 6 co @ 2.4 GHz E5-2440 totaling 12 cores
128 GB Ram
vSphere Enterprise Plus
Assume Linux – pick your distro of choice
3 year support
No internal HDD - assuming USB boot
2 x dual port 10 GbE
2 x dual port Fibre
All power cores, rail kit, misc
$25,183 each server List price
.
.
Oracle Licensing cost
Enterprise Edition - $47,500 per core
EE maintenance @ 22% per year - $10,450 / co
RAC - $23,000 per core
RAC maintenance @ 22% per year - $5,060 / co
.
.
IBM’s latest Power8 server
S824 Power8 server - qty 1
(this is the 2 socket model although I just selected it with 1 socket)
.
8 x 4.15 Ghz Power8 cores
256 GB Ram (need more Ram because I’m planning to host more VM’s)
DVD
Split backplane
4 x SSD (building the way I would built it and not just to lower the price which I could do by using HDD)
2 x dual port 10 Gbe adapters
2 x dual port Fibre adapters
AIX v7.1
PowerVM Enterprise Edition
3 year 24 x 7 maintenance
$79,807 server list price
vMotion from a Dell to a Cisco then? How about from a Westmere to a Ivy Bridge EX based server? How many vCPU's do you have in this 1 TB VM by the way? With just 2 HT how does that scale? I'm not trying to pick but just point out that just because it works in a lab doesn't mean it will work in a customer environment under load using a shared network. By the way, these features I describe do scale and do work with Power8 and PowerVM from min to max of it's capabilities.
With regard to your statement about "Licensing is not an issue". Licensing is always an issue on x86.The server are bigger because Intel creates bigger chips that are weaker per core but greater by their overall sum than the previous sum of the socket. Looks like a win but when a lot of software is priced by core like Oracle, it very much matters. Oracle DOES NOT acknowledge VMware for sub-capacity licensing purposes. You cannot turn off cores and you must license all cores in the server times .5 (ie 1/2). 60 core Ivy Bridge is 30 Oracle licenses even if running at 20% utilization with or without VMware. At $47,500 per core times 30 cores I would argue that licensing very much matters. You make the point for why businesses should consolidate x86 workloads onto servers like Power8....I wish I could also say Itanium and SPARC but I just don't see their technology keeping pace despite the claims. Would your employer accept you working at 20% utilization? Then why accept it for the servers? Take 5 of those 20% servers and consolidate onto a single server...better yet, take 10 or 20 of them because like rocks, pebbles and sand fill up a jar so do more workloads fit onto a (Power) processor and it's scheduler to drive more efficiency.
It's easy to say "Look what we have". Show some benchmarks. Run two systems side by side. One with Hana and no VMware and another with VMware and let's see how they compare. Further, what is the benefit of adding the cost and performance penalty of VMware? To use VMotion? Should be interesting to watch it try to move 1 TB of "in-memory" data.
I think customers are better off running their SAP environments using AIX on Power8 with DB2 10.5 with BLU Acceleration. They can virtualize everything, get the reliability and scalability and with Power8 they can get 2X the performance over x86 not to mention what DB2 BLU can do....wow - the compression, speed and the fact you don't have to buy huge memory machines to do it all. Last report I saw comparing DB2 vs Hana showed it was quite a bit less expensive as well.
@MadMike, only because you've pushed my buttons with so many ridiculous statements that I am taking the time to respond. You throw crap into the wind hoping it blows away from you. This time it is coming right back onto you. Solaris is not the hottest OS, as another gentlemen said that would be Linux in general and AIX in the Unix space. Your facts are wrong about Power. A 795 with 256 cores only needs about 64 to outperform a 256 core M9000. Any of the Power7 servers require 1/2 the cores on a staightline race to beat T5, factor in the efficiency of the hypervisor running dozens of VM's and it goes up to 10 - 20X. Just so you don't cherry pick a benchmark and say "look here". You pick your T5 server and I'll bring my largest Power8 server which is 24 cores and let's have at it. Let's run'em to 100% utilization both running Oracle then I'll do another run with DB2 10.5 just to show how much further it scales than Oracle. Did I mention I will be bringing my FlashSystem 840 - don't worry about that though - you will have your 128 core T5 right! By the way, just to show you were wrong, a 64 core 770 supports 1000 VM's (IBM artificially limits it at this number) and can create up to 5120 threads in the server if needed but who would. The point is, you don't know what you are talking about let alone knowing anything about Power technology. By your logic since IBM invented DRAM and HDD, Oracle is "copying" them? How childish! AIX WPAR's are not copied from Solaris Containers. Maybe Containers came from BSD Jails?
With regard to AIX being killed off - what are you smoking? IBM's announcement of Power8 on April 28th supports all 3 of it's core OSes as it has for a decade - IBM i, Linux and AIX. It is as committed to AIX and IBM i as it ever has been. You are hearing about Linux so much because IBM is trying to grow this market. It owns the Unix market with AIX. No matter how much they were to market AIX and PowerVM to some in the open source community they won't budge, so unlike what Sun and now Oracle couldn't do which is embrace open without conditions (ie CDDL), IBM created OpenPower Foundation, is using OpenStack for systems management, added PowerKVM as a option for Linux users in lieu of PowerVM, and added Canonical Ubuntu and Debian Linux to the stable with RedHat and SuSE. With Power8, it runs Java 2X over x86 and Linux (initially Ubuntu) in either Little Endian or Big Endian mode (ummm T5, what say you?) All it takes for LE support is a recompile. (read as period). Ask Ubuntu how effortless it was for them. Take a look at http://benchmarkingblog.wordpress.com/ to see the new #1 benchmarks from these 2 socket servers for SAP, Oracle, Siebel, Java, SPECfp and SPECint. Of course, unlike Oracle's benchmarks where they beat and compare to themselves you may not be that impressed. Last comment. In 2015, Oracle will release a ..... yeah, because they are so reliable in delivering what they say they will. I predict you will see a 2 socket Power8 server with more memory before you will see a SPARC server with 64 TB of RAM. Get back to your basement and clean your room!
Are you confused about product positioning, features, etc. I was in STG with Power. This is no different than Lexus vs Toyota vs Scion. or Cadillac vs Buick vs Saturn - pick your analogy. Power & z didn't fight. Power was and is thankful for Z as the platform owes a lot of its technology to what the ultimate compute platform offers. Sounds like you have the same Napoleon complex the other x86 vendors have with their virtualization technologies - we can do what the big boys do and our servers cost $10K. The reality is you are letting your emotions and fondness for your preferred platform make you sensitive to product positioning. Z servers have their place in the data center along with Power just like x86. x86 vendors are trying to grow their marketshare by saying more cores and more memory makes a enterprise server.
I am no longer at IBM and instead a reseller. If you were sensitive to what you heard in STG from Power sellers then you won't be happy with Power8 because that platform is bringing it's enterprise features of reliability, scalability, security, virtualization and flexibility with AIX and Ubuntu, SuSE and Redhat Linux running Linux apps in LE or BE mode. Port and compile or just compile and run to get 2X the performance and utilization levels up to 90%. Scale out or scale up with Power8 at x86 price parity, consolidate dozens of x86 at x86 price comparable or go with enterprise Power servers to run mission critical workloads 24x7x365....and yes, if customers need fault tolerance with the ultimate in reliability and serviceability then they should go with a mainframe.
There was no Anti-x86 faction in STG. There were 4 distinct products in STG - z, p, storage and x.They are all distinct. Two problems I've observed is IBM won't make a inferior server product if they can help it. Thus, the X5 and X6 are top of their class. However, the market doesn't buy x86 for it's technology they would rather buy it for its price point. So, a highly reliable X6 3850 is 30% higher than a commodity x86 server from some vanilla vendor. Secondly, IBM made/makes it a challenge to sell their x86 servers compared to other vendors. Other vendors tear down obstacles whereas it seems like IBM puts them up. It's a shame because from a technology perspective it is outstanding in the x86 space. It's not Power but I would select it for the right workload.
@obnoxiousGit well suits you as its clear you are a contrarian who likes to argue for arguments sake. Looking for gotchas and wild goose chasing challenges. There have been plenty of reasons provided which you fail to accept, You can make all of the noise you want if that is what fulfills you. Customers who real problems though are not looking to follow the sheep. That is what got them where they are at now. Scale out with great when Unix was big iron and expensive and all they had were 2, 4 & 8 core servers doing file and print services or low end SQL server environments. Just because the compute is now 15 co per socket doesn't mean the server is capable of 2-8X the workloads. Server has to be reliable. There are vast differences in this area among the various vendors. Two vendors I have seen which standout are IBM and Fujitsu - both make a quality x86 server with enterprise RAS features. There are other very popular vendors that have vanilla RAS features who focus on commodity price focused servers and no technology. I wouldn't replace a 795 with 1 x86...I would replace a 40 core x86 with a 4 core Power server. I would replace 500 x86 servers with a single 795. Move on!
System z and Power share IP so even though one is CISC and the other RISC there is quite a bit in common - which is great for Power consumers! In order to see the value, one has to open their eyes. Power delivers and has delivered a reliable platform for years unlike x86 which is getting better but the key word is "getting". Layer on top the secure, no CVE PowerVM virtualization that should be called "Flexible" as you can have it your way unlike the leading x86 offering supporting the 2 thread only servers. Threads are on or off unlike Power which can have 1, 2, 4 or 8 set separately in each VM for any workload type further improving the financial justification. x86 doesn't satisfy 95% of the needs as that implies 95% of the users are satisfied. I would argue the constant instability, security issues, patch Mondays and the answer to everything is "just buy another cheap server" with no regard for the actual cost of what it means to add the 200th server to the environment when those 200 might consolidate onto a single Power server. It's not that x86 dominates rather than 95% are misinformed of the changes Power7 brought customers as IBM changed directions to battle x86 and the major OS and virtualization players. With Power8 the value proposition is stronger and if you think x86 is everywhere then your eyes are wide shut as Power processors are prevalent throughout our lives starting with your vehicle.
Having worked at Sun for 10 years I have my own perspective. They embraced "open" begrudgingly as a defensive measure. When they did, they restricted and locked down the Sun Linux so it was unusable. It was all in a effort to protect their profitable high end SPARC business. To IBM's credit they are not afraid to turn the ship into the wind and take on the competition. As a IBM BP I can tell you it makes it more challenging for us but at the same time I have the choice to embrace outstanding technology and go out and compete OR feel sorry for myself and simply ride it out and jump on another bandwagon. I like the competition so I'll compete! Go Power! Power has a offering that has price parity with x86 in the scale out space as well as the traditional high end servers to meet the most demanding customer needs.
It is about reliability, serviceability and operational benefits. What platform was Target running in each store? What virtualization product leads the pack in CVE's? Are you in favor of any OS on x86 or just one, say Linux? If so, is Linux the best OS because it runs on x86 or does it make x86 better? If you like Linux and it makes x86 better then I would argue that on Power you would love it because Power makes Linux even better. It brings open source as close to a commercial Unix OS possible while still maintaining all of the open source characteristics. Lastly, even if Power8 didn't outperform x86, it comes down to TCO. TCO is made up of software licensing, software maintenance, # of FTE's required to support the solution, # of copies of virtualization & OS and all of the other software required. When you can run 1 thread in 1 VM, 2 threads in another VM, 4 threads in another VM and 8 threads in another VM all with a mix of dedicated or virtual I/O then Power brings a reduced footprint which can drive the processor utilization up to 90% delivering quality of service and availability to each VM at massive software savings.
IBM brings innovations, technology, competition and the only alternative to a platform that must be cobbled together with expensive software and other unnatural means to overcome its many deficiencies. With these announcements customers now can choose to run commercial IBM i applications in a integrated stack, scalable Unix applications with AIX and now choice of Linux distro's with Ubuntu, SuSE and RedHat support LE or BE all managed using a OpenStack based solution. All of this AND it comes on the most secure virtualization platform outside of the mainframe where the answer is Yes to your question if you can do something with PowerVM vs the alternatives. Finish it off with performance that bests both the Power7+ servers it replaces and is more than a leap frog ahead of x86 but some serious distance. That isn't about bragging, that is about reducing software licensing cost.
What would you rather have? One chipset, one virtualization offering, one OS, one storage vendor, one DB.....? Do you want one TV channel? Cars only in one color? Come on, competition breeds innovation - you should applaud this and be glad to see it because if it doesn't make the x86 vendors better it will make the ARM vendors better.
Actually, if you read my first postings I gave credit to several vendors such as Pure Storage and HDS. They do things differently but viable options. Even though I'm not a fan of EMC you cannot overlook their place in the industry or success and I said as much. I think you confused me with the "Anonymous Coward" posting. I am comfortable in your knowing who I am and what I do. I stand behind what I say. Admittedly, I make mistakes and have opportunities to learn and grow - that is life.
If we had the benefit of a whiteboard I would show you the benefits and ease of it all. I approach solutions looking to provide the following: reliability, flexibility, scalability and performance at the desired price point just like we all look for as consumers. Unlike many vendors who try to lock in customers to products I think we should give them maximum flexibility. When we do, customers embrace both the technology and the seller for having the credibility to do so. It is too bad if you have been burned by IBM or some overzealous sales people. I am a technical seller and manage our Power practice at a IBM partner. I'm here because I love my job - like a pro athlete who gets paid. If I didn't I would go elsewhere.
@Matt_Bryant I think you are the pot calling the kettle black. I am not a drone but a fan of technology that works and delivers value. Products I believe in and that solve problems. When understood and not using superficial comparisons to the competition like a x86 server is the same as a Power server or VMware is the same as PowerVM/Power Hypervisor. ie a server is a server and virtualization is virtualization. When you know how to exploit Power servers, hypervisor and thus virtualization and storage like the 840 then you see how a workload can do more with (far) less, faster and more reliably. So Matt_Bryant, you can attack me if you like as I have watched your behavior on these forums for years and that is your MO. The facts are what they are and sometimes when people speak them they are truth and not marketing speak or bile as you so eloquently state.
The cost of IBM’s FlashSystems is such that customers should now start considering buying it as primary storage. Why would you not size for capacity? When a 1U, 30 lb, 300 Watt, 24 TB, 25 microsecond, 500K+ IOPS FlashSystem 820 has as many IOPS as ~1550 HDD’s in a EMC vMAX. To the comment about cost of SVC + IBM Flash what do you think is cheaper – a few RU of Flash sized to the capacity needed behind a virtualization engine delivering the ACS? Eliminate I/O from the equation. SVC adds around 100 microseconds to each I/O. When you have products with latency as low as 15 and 25 microseconds adding 100 us really isn’t bad now is it – oh my, my latency is 150 or even 250 us. You have to give EMC, HP and Dell something to criticize since they typically claim latency <1ms. Hmm, if the Chevy Camaro could produce 600 HP or more like the Ford Mustang don’t you think they would claim it? When they can’t they say “who really needs that much HP when the speed limit is just 55 mph”. Bringing it back to storage they will claim that customers just want to use flash for tiering – yeah right, because that is all your storage is really good at.
To finish up on the 840. Josh Goldstein doesn’t know what he is talking about. I have 12 significant enhancements tohe 840 which is an enterprise product offering up to 48 TB in 2U at roughly 75 lb and 625 Watts. It is fully hot swappable with no single point of failure. It has up to 16 Fibre ports and supports FC, FCoE and IB. That capacity equates to 1 PB of pure Beast in a 42U rack. After racking, to begin the configuration until you have storage presented for use it will take less than 5 minutes. Augment existing SAN storage such as HDS, EMC, NetApp, Compellent or IBM with a 820 or 840. Zone a portion to use for enhanced Read performance. Virtualize the rest and present storage to databases, logs, VDI – everything because it is less expensive and the numbers make sense when customers put their Ford vs Chevy biases down to hear the message.
The article fails to capture the essence of IBM’s FlashSystems products made up of x10, x20 and 840 models where x = 7 or 8. These products are the “Beast” or the 600 HP engine in a Ford Mustang. You enable all of the Advanced Copy Services (ACS) or the “Beauty” when you pair it with a SVC, V7000 or V5000 products. This gives you “Beauty and the Beast” – Clever huh? I thought of it myself ☺. The article presents the idea of SVC as simply adding cost. You need to take a look again at what SVC is. Unlike EMC’s vPlex which is about locking you in to there products, SVC frees you from all storage vendors. It is storage agnostic. Manage virtually any storage that is Fibre Channel – IBM or not. Manage just a portion of it. Bring in a SVC with a storage refresh and place the old storage behind the SVC and now you have flexibility to not only move data from old to new but also you can extend the life of that old storage. If the SVC which is an enterprise class product is too expensive then look at the V5000 or V7000 which is mid-range. Put the minimum number of disks in and you can still manage external storage to include FlashSystems.
What would you expect them to say about any IBM product? Excuse me Mr Honda seller - how does your car compare to the Toyota? This is a ridiculous article. What it does do is show who has a weak flash offering and who respectable competitive offering. Start with the respectable and competitive vendors which are HDS and Pure Storage. They have their approach which is fine "6 to one or half dozen to another". However, take EMC, HP and Dell, which do not have viable or scalable flash products. EMC has XtremeIO which calls a storage shelf a "brick" - that should say it all. Doesn't integrate with the rest of their stack and is the Yugo of their product line. Step into their class leading midrange and high end and if you want flash you had better open up Ft Knox. They use form factor SSD which is flash, they call it flash but everybody knows and admits (except those who don't have other options) it is encumbered by the form factor, storage subsystem, etc unlike IBM's FlashSystems which uses DIMM style Flash Modules that are purpose built. This EMC SSD flash can cost you upwards of $75K per TB for the mid-range & high-end. Guess who loves their flash?!
Good old HP and their 3PAR 7450 is yet another form factor SSD flash based product that uses 3.5” SLC based or 2.5” MLC based SSD. MLC or consumer grade flash which is cheap for them to produce presumably to drive their profits while selling you a swan song of performance putting your data at risk. A simply Google search on MLC vs eMLC vs SLC will show where MLC rates significantly lower than the others. People are moving away from SLC and settling on eMLC as the balance of performance, reliability, density and cost. And Dell – the spokesperson who said when our RoCE product comes out in a few months it will kick the 840’s butt. IBM is coming out with the FlashSystem 880 soon that will have 1 billion IOPS with latency as low as 25 nanoseconds – trust me, I said “soon”.
This will be a long winded view of what I see happening that differs from many of you. This announcement is a good thing for IBM, Lenovo and customers. Lenovo, who is the #1 global PC vendor clearly wants to move into the enterprise with a quality x86 kit that includes converged capabilities. IBM System x and Flex is the obvious platform. They pick up a technology rich X6 platform along with solid entry level 2 socket sockets servers. They also pick up the extremely flexible Flex chassis that is truly converged in within the chassis unlike the competitors which are merely an assembly of best of breed by their own definition. I'm not selling the features, just stating the facts.
I do not have all of the details but this announcements is a win-win-win as IBM will pick up the supply chain efficiencies of Lenovo to lower the cost of manufacturing and selling Flex chassis and x86 nodes. I also imagine there is engineering / manufacturing agreements between the two that share R&D & costs as they jointly work to improve the platform for future models. With Lenovo agreeing to license existing IBM technologies like storage and software into their markets, IBM selling it's x86 business and thus the financial burden to Lenovo while still benefiting from the platform via Lenovo efficiencies this should make Lenovo more formidable against Dell and HP while strengthening IBM's balance sheet.
The point is that the market for System x just grew from the IBM x86 sellers to the #1 global PC vendor's market. A few of those new customers will get introduced to IBM storage, software and hear the Power server story. The Flex technology is solid, Lenovo is a supply chain monster with huge global markets, IBM just shifted some of its cost to a partner while pocketing some cash it looks to me like they are crazy like a fox.
Excellent comments as you have captured exactly the essence of the big picture. Something this author and others who I have read from Reuters, WSJ and others after the earnings report do not seem to realize. They seem to think the only revenue earned from the initial product sale is what goes out the door. When I read the article my first thought was to look for the tag "Sponsored by EMC" or pick another competitor as they are the ones who would want to see IBM out of the storage, server, software and services business. I do not work for IBM but work with their products as well as products from other vendors. IBM has it's warts like any company but they are a true engineering and technology company. Their storage is very, very good. The DS8x00 family is as solid as anything from EMC or Hitachi. XIV bridges a gap the others have a hard time matching up against. V7000/V5000/V37000 are all variations of a excellent recipe and very competitive against EMC's VNX. It can become a Chevy vs Ford discussion at times on features by by and large when you can have a fair fight I see the IBM technology stack outperform EMC more times than not. I see EMC sales tactics outperform IBM more times than not - just my observations so don't take it as an indictment, just an observation. IBM's server technology whether Power or x86 is vastly superior to anything from HP, Dell, Oracle (considering RISC & x86 if they offer it). It comes back to the IP in the servers which is amazing. That delivers servers that are more reliable, efficient, scalable, flexible, often higher performing and on occasion more costly - depends on what is being compared. However, it becomes relative because the technology, especially with Power is so efficient that it takes far, far fewer resources to do the job than other platforms just smaller servers, fewer software licenses and thus the TCA and TCO gives the advantage to IBM. I could go on. Point is that IBM innovates and sells. When they sell they try to sell additional features to grow the ecosystem. Unlike Microsoft which requires a 3rd party company to product and sell a x86 computer to get their OS into a household where they then try to sell other products like MS Office. IBM tries to sell the hardware, OS, Application, services and support.
Clearly the Oracle marketing team had an extra shot of espresso today. Statements about SPARC outperforming Power7/7+, copying SPARC features, SPARC is significantly less expensive than Power, oh yal and the old reliable "Oracle software licensing cost is 2X more on Power, etc, etc is laughable. These statements are misleading at best.
You compare your Sept '13 T5 with 128 cores to a 2008 entry for a Power6 (N-2 technology) with 64 cores and only yield a 40% better number.
Exa* is not best of breed, rather it is cheap white boxes cobbled together with expensive Oracle RAC clustering, limited SAN support, internal flash, etc. If it was best of breed it would use Cisco, HP or IBM x86 technology with VMware and flash technology from Fusion, Violin or IBM Flashsystems (ie TMS)....oh, by the way then it would also use IBM DB2 and pureScale instead of Oracle Database and RAC.
The net net is that benchmarks are great and fun to watch. All vendors should do them....nice to see Snoracle finally do some as they have been largely absent for years. However, it really comes down to the efficiency of the server to deliver that performance to real life workloads. I just read how the SPARC M6 uses critical threads to essentially hog all of the processor resources. That isn't so good for running multiple VM's on a server is it? With Power servers I can deliver a QoS to each VM as well as dynamically change how much cpu, memory and I/O to increase performance, throughput or both based on the businesses need.
Oh yeah, lastly to the old reliable "Power is twice as expensive as x86 or SPARC" comment. There is more to the story than just the Oracle licensing factor with Power being 1.0 and most other platforms being at 0.5. Because of the efficiency (and performance) of a Power server I can usually run a workload with 1/10th the number of resources required by x86 or SPARC. So, even with Oracle manipulating the licensing factors that still means a x86 or SPARC server would require 5X more licenses.
I'm in the channel and experiencing some of this change first hand. I can tell you that I have never seen IBM as focused or agile in making decisions as they are right now. From Doug down to the IBM direct sellers and business partners there is a energy and excitement for the sellers to take a rich and robust Power platform and compete against and in the new competitive marketplace which now includes Linux, x86 and VMware.
Power servers are about "balanced" computing. There is a balance between the cores, memory and I/O. This is the reason a 16 core server can often (and regularly does) do the workload of a 64 core x86, Itanium or SPARC server. Power memory doesn't need to run at 1600 MHz like x86. - this is an example of the shiny object that distracts consumers to buy off the spec sheet.
Power RAS capabilities far exceed x86 in every case except for the very high end x86 servers. These high end x86 servers differ from the standard 2 & 4 socket x86 servers.
I'm rambling - I over explain stuff but try to avoid confusion. Regards
snipped by mod for brevity
Power servers deliver more value than any x86 server on the market not to mention all other Unix offerings. Power's value prop is that it delivers performance, reliability, flexibility and efficiency at competitive prices. Long gone are the Big Iron days. Today, with IBM's Power7+ lineup they have 1, 2 and 4 socket servers that are price comparable to x86 for a similar configuration. Thankfully IBM continues to deliver a premium product to meet customers needs at varying levels. Similar to car manufacturers that offer sub-compact, compact, mid-size, large and luxury cars with varying features and price points IBM has the same in support of IBM i (good 'ol 400), AIX and Linux on servers ranging from 4 - 256 cores running at speeds up to 4.42 GHz and memory capacities from 4 - 16 TB.
Many if not most x86 customers have been trained to buy commodity servers lacking in reliability features requiring horizontal scaling in data centers. Not a big deal until you run out of power, cooling and space not to mention the FTE's required to support that. It's interesting to read the comment earlier posted about the cost of a Unix FTE costing more. I say UNIX because HP-UX, Solaris and AIX are all quality operating systems and it is often the case their platforms run a large majority of a businesses operations. I have a customer that runs 80% of their business on Power with 2 FTE's and the 20% balance is using x86 that requires 6 FTE's. This is attributed to the inherent reliability of the OS and server not to mention their ability to efficiently run significantly larger and more workloads.
IBM could play the "I'm cheapest" game like I guess Dell and SuperMicro do as stated above. If that works for your organization that is great. However, Power servers are price competitive. Where x86 requires 2, 4 or more servers for many workloads often requiring expensive clustering software you often only need 1 Power server. Some businesses demand a cluster as part of their risk mitigation strategy and with Power servers they only need two servers then. Not because the servers aren't reliable but because the business dictates that reasonable measures to avoid outages be taken. Because fewer Power servers are required the TCA and TCO continue to favor Power.
Lastly for this comment you should know that Power servers not only generally outperform x86 servers on a core for core basis but that takes a back seat to their real value which is their efficiency. For customers it isn't about the price of a $15k x86 servers when you have to buy 4 to do the job of 1 Power server. The software on those 4 servers may cost $1M compared to $250k on the single Power server. Or, in the case of EnterpriseDB there is just the support cost on a 8, 16 or 32 core Power server that I would put up against 64 - 128...even 256 x86 cores to do the same work....I've even seen it take 300 x86 (latest x86 chipsets) cores that only take 30 Power cores to do. This isn't hyperbole it is what Power servers are - inherently reliable, flexible and efficient.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
