Friday, December 26, 2008
"Final (2008) score: 4 right and 11 wrong. ... Obviously I have to start making vaguer predictions ..." :-)
Despite this self-flagellation, he's generally had some pretty good insights prior to 2008 and you can't win 'em all, all the time. Here are some of his darts-in-the-wall for 2009:
- "The good news is that most recessions mean new IT platforms."
- "The next Yahoo CEO will dismember the company and sell it piecemeal."
- "Microsoft and Google will peak in 2009 ... Steve Ballmer is toast." [ Jenks claims 2010 for the M$ demise.---njg] (Like Humpty-Dumpty, The Goog) "... is too fat and happy."
- "If Microsoft and Google are down then what's up? Apple!"
Please standby. ...
Friday, December 19, 2008
The role of the Guerrilla Manifesto is to provide you with more authoritative support via the various mantras listed there. Hardcopy is good, but online is better; especially for use with your mobile phone. The most recent updates are now available and have been indexed for easier reference. Let me know how this works for you.
If you didn't attend CMG 2008, but you'd like to know more about capacity management techniques, consider coming to my 2-day Guerrilla Boot Camp class in 2009.
Tuesday, December 16, 2008
Tuesday, December 2, 2008
A team of researchers from ETH Zürich (Einstein's alma mater) in Switzerland used four decades of data detailing the evolution of open source software applications created for a Linux operating system to confirm the adherence to Zipf's law. The team studied Debian Linux, as it is continuously being developed by more than 1,000 volunteers from around the world. Beginning with 474 packages in 1996, Debian has expanded to include more than 18,000 packages today. The packages form an intricate network, with some packages having greater connectivity than others, as defined by how many other packages depend on a given package.
It's also amazing what passes for physics these days:
T. Maillart.; D. Sornette; S. Spaeth, and G. von Krogh. “Empircal Tests of Zipf’s Law Mechanism in Open Source Linux Distribution.” Physical Review Letters 101 218701 (2008).
Wednesday, November 26, 2008
Monday, November 24, 2008
The breakdown of these top-100 machines by processor family (not shown there) looks like this:
- Intel: 75.6%
- IBM: 12%
- AMD: 12%
- NEC: 0.2%
- SPARC: 0.2%
Wednesday, November 19, 2008
A conceptual difficulty can arise when we try to summarize a set of performance numbers as a single number; especially if they're rates.
Thursday, October 23, 2008
Wednesday, October 22, 2008
Wednesday, October 15, 2008
Tuesday, October 14, 2008
Put differently, the less you know about the system, the more inclined you are to see patterns that aren't there or that aren't meaningful. This is also one of the potential pitfalls of relying on sophisticated data-visualization tools. The more sophisticated the tools, the more likely you are to be seduced into believing that any observed patterns are meaningful. As I've said elsewhere ...
The research experiments used very grainy pictures, some of which had embedded images and others which did not.
Friday, October 10, 2008
Naturally, they commence by pointing fingers at Wall St. (hint, hint, hint) and the lack of regulatory oversight (although, I think Sequoia is probably more concerned about potential new regulations). Paraphrasing some of their points that are relevant to the IT sector:
- In terms of the standard econo-metrics, this crisis appears to be unique and therefore recovery may take much longer.
- If you want to see the future, look at Japan's recent economic history.
- Manage cash expenditures and focus on quality. [That last one's a doozy. QUALITY ... What's that!?--ed.]
- Advertising markets are cracking. Retail and e-commerce are deteriorating. Mobile is not immune.
- IT spending is now being scrutinized more than at any time since the 2003-2007 period. [??? But 2007 was only last year.--ed.]
- These concerns have triggered a sudden and unexpected drop in business activity.
Although these comments are based on US market data, we do live in a global village . Here is the Sequoia presentation.
Thursday, September 25, 2008
Click on the image for details.
For those of you coming from international locations, here is a table of current exchange rates. We look forward to seeing all of you in November!
Thursday, September 18, 2008
Wednesday, September 10, 2008
"Traders experienced problems connecting to TradElect, a 15-month-old proprietary LSE platform developed with Microsoft Corp. technology that the LSE has touted as allowing it to expand and speed up its capacity for trades."
"By early afternoon, the LSE was informally telling customers that its system would be back online. That appears to have created a logjam as customers tried to reconnect at the same time."
This is almost a replica of a problem described by my colleague Steve Jenkin in our 2006 CMG-Australia paper:
"That load was completely disabled to improve system throughput for the most interesting capacity planning period on the busiest day. This scheduling was mandated by the Secure Hosting Facility after a software fault caused a flood of more than 6,000 emails to be issued after a multi-day stoppage. Such an email flood, containing several days of traffic in 10 minutes, brought the facility mail system to its knees ..."
Some things ...
Friday, September 5, 2008
This early review indicates the luster is already off Chrome, due to some serious performance issues related to the Adobe Flash plug-in being a CPU hog. Unlike FireFox, there's no workaround.
Thursday, September 4, 2008
Guerrilla graduate Scott J., sent me the following cartoon by Scott Adams, which offers a Dilbert profundity on the same topic.
For me, the punch line IS the 2nd frame. You may need to click on the strip to see the Dilbert punch line in the 3rd frame.
Wednesday, September 3, 2008
If you're in the area, you might also pass this on to other interested colleagues and otherwise contact Anne-Marie Eliseo at email@example.com or phone: +61-8-8302-3928 for more details. They are trying to figure out if they should run it in Sydney as well. Your feedback could help them.
Wednesday, August 27, 2008
"That all sorts of forces, from Wall Street's quarterly profit pressures to worsening government deficits, have undermined support for the forward-looking research on which future products will depend."
I've certainly been complaining for a long time about the excessive influence of Wall St. on shortening capacity planning horizons. Estrin also claims that the environment for innovation in Silicon Valley and the United States has already changed for the worse. Whether you agree everything she says or not, I think her comments are important because there seem to be precious few figureheads who are prepared to stand up and tell as it is.
Updated on Thu Aug 28, 2008: Case in point from Slashdot regarding Alcatel-Lucent (née Bell Labs) yanking the financial plug on basic science "like yer startin' a mower!" (to quote Elaine Benes). As I've said before , the Edison model is dead.
Saturday, August 23, 2008
Thursday, August 21, 2008
Back in April, I was busily directing the attention of the GCaP course attendees to the Guerrilla Manual booklet inserted in the back cover when, to their surprise and my chagrin, they discovered that their copies did not have the booklet. Mine did, but it was an older copy. What the ...!? Turns out, it was a production error in the latest printing, which is now in the process of being corrected. The misprint versions have possibly been "recalled" and that's why it is temporarily out of stock. I'm sure Amazon will still be happy to take your order. Thanks, Greg.
Tuesday, August 12, 2008
Tuesday, July 29, 2008
Friday, July 18, 2008
- Simulations that supported the theorem involved a barrier synchronizer, and for this to work continuously, my colleague Jim Holtman, discovered that the distribution of parallel execution times (with mean 'think' time Z) could only be deterministic (D). That's a special case of the repairman: D/M/1//N, whereas the equations in my proof show quite clearly that it must work for M/M/1//N.
- There's a theorem in queueing theory (Bunday and Scraton, 1980) which states that what holds for the repairman (M/M/1//N) also holds for a more general source of requests viz., G/M/1//N. Since G is a superset of the distributions D and M, this theorem suggests that our simulations probably shouldn't be restricted to D-distributed think/execution times as observed in (1). Or should they?
Monday, July 14, 2008
In this class, computer engineering and statistics expert Prof. David Lilja presents an easy introduction to statistical methods and finally leads us into the topic of Design of Experiment (DOE) methods applied to performance and capacity planning data.
Having established the foundation theory, R expert, Jim Holtman will show you how to apply DOE and other statistical techniques using example case studies.
You can register for the class, and book your hotel room, online. Book early, book often! We look forward to seeing you in August.
Saturday, July 12, 2008
In the aftermath, Apple managed to sell 1 million iPhones in just 3 days cf. the same number of the original iPhone in 2.5 months. Why should you care? Mobile devices like this are actually computers, not just phones. The iPhone runs MacOS X; same as mac laptops. In my view, this and similar devices represent the commodity computer of the future.
Wednesday, July 9, 2008
I spent part of last month in eastern Germany (former DDR) where there is a decided lack of internet connectivity (hence, no blog entries), even after west Germany (FDR) has thrown a considerable amount of euros to bring the eastern economy up to speed. Hotels either don't offer internet at all or want an arm and a leg for it, if they do.
As far as I can tell, eastern hotels (most of which really don't have the inclination to understand IT; just like hotels in the USA) have let Deutsche Telecom swoop in monopolitically to set up internet packages whereby they can charge 9 euros (a whopping $14 USD) for just 1 hour online, unless you are a card-carrying DT customer, in which case you are entitled to some kind of discount. After resisting for days, I eventually broke down just to prevent my web mailbox from overflowing. These packages are affectionately termed "Hot Spots"; a term possibly meant to confuse tourists like me who naturally jump to the conclusion that it means New York style free wi-fi. It's wi-fi ("WLan") all right, but at a very hot price.
On the FOSS scene, we have VirtualBox as a replacement for VMWare Fusion or Parallels.
Saturday, May 24, 2008
- German: Linux könnte seine Position im Servermarkt ausbauen, wenn es dem Vorbild der Mainframes folgte und deren raffiniertes Performance-Management übernähme.
- English: Linux could be in a position to expand its presence in the server market by looking to mainframe computer performance management as a role model and adapting its instrumentation accordingly.
Topics discussed include a comparison of time-share scheduling (TSS) with fair-share scheduling (FSS) and the Linux Completely fair scheduler (CFS), how to achieve a more uniform interface to performance and capacity planning measurements, and the kind of advanced system management capabilities available on IBM System Z for both their mainframes and clusters.
Thursday, May 15, 2008
Wednesday, May 14, 2008
Monday, May 12, 2008
Climatologists, up the road here at LBL, claim that a supercomputer using about 20 million embedded microprocessors, such as those found in cellphones, iPods, and other consumer electronic devices, would deliver useful climate simulation results at a cost of $75 million to construct. A comparable supercomputer could cost closer to $1 billion. Based on a recent post, I'd be wanting to see the front-end compiler system that can upload 20 million processors.
Wednesday, May 7, 2008
For a long time, I've thought it would be cool to be able to visualize system performance as a shape but was never quite sure what that meant. My role model has been SciViz, where complicated system dynamics like the time-development of tornadoes can be visualized in 3D animations. More recently, the cyclone paradigm has been used for textual analysis based on word repetition (The novel "Frankenstein" is show above). The more a word is used, the larger is its cube. Blue cubes are words that are unique, red cubes are not. The diameter of the rings is determined by the size of the paragraphs. Who woulda thunk it?
The closest I've come to producing performance data as a "shape" is this:
which shows processor %user, %system, and %idle time for a 72-way SMP running a transaction workload on ORACLE 10g over a 20 minute measurement period. Data supplied by Tim Cook of Sun Microsystems. The time-development of the data (not shown here) is not too far removed from the tornado animation in the first figure.
Monday, May 5, 2008
Supply Chain Factoid: There are so many more (full) shipping containers coming from Asia to the USA and Europe than going the other way, that it is less cost-effective to store the empties than to simply scrap them and make new containers as needed.
Saturday, May 3, 2008
Friday, May 2, 2008
Tim McCluskey wrote on 4/30/08 12:29 PM:
The class was great! I'm glad that I didn't listen to the 2 people that gave me the impression that it or you were going to be over my head.
Vladimir Begun wrote on 4/30/08 20:33:36 -0700:
... it was great! A clear and horizon-expanding presentation of an actual experience in the capacity planning. About right for the jump-start! Eager to attend the level II class.
More people enrolled than we had hoped for in the shortened time-frame that was available to advertise it, and they all reported liking the hotel, sleeping rooms, seminar room, food, and especially the free wifi everywhere. They were also grateful for being able to use the ensuite bathroom across the hall instead of having to walk to the other end of the hotel corridor. :-)
We'll do it all again, in a couple of weeks but at Level II, this time.
As I discuss in Chap. 9 of my Perl::PDQ book, the qualitative design rules for incrementally increasing the scalability of distributed apps go something like this:
- Move code, e,g., business logic, from the App Servers to the RDBMS backend and repartition the DB tables.
- Use load balancers between every tier. This step can accommodate multi-millions of pageviews per day.
- Partition users into groups of 100,000 across replicated clusters with partitioned DB tables.
All the big web sites (e.g., eBay.com and EA.com) do this kind of thing. But these rules-of-thumb beg the question, How can I quantify the expected performance improvement for each of these steps? Here, I only hear silence. But there is an answer: the Universal Scalability Law. However, it needs to be generalized to accommodate the concept of homogeneous clustering, and I do just that in Section 4.6 of my GCaP book.
The following slides (from 2001) give the basic idea from the standpoint of hardware scalability.
Think of each box as an SMP containing p-processors or a CMP with p-cores. These processors are connected by a local bus, e.g., a shared-memory bus; the intra-node bus. Therefore, we can apply the Universal Scalability model as usual, keeping in mind that the 2 model parameters refer to local effects only. The data for determining those parameters via regression could come from workload simulation tools like LoadRunner. To quantify what happens in a cluster with k-nodes, an equivalent set of measurements have to be made using the global interconnect between cluster nodes; the inter-node bus. Applying the same statistical regression technique to those data gives a corresponding pair of parameters for global scalability.
The generalized scalability law for clusters is shown in the 5th slide. If (in some perfect world) all the overheads were exactly zero, then the clusters would scale linearly (slide 6). However, things get more interesting in the real world because the scaling curves can cross each other in unintuitive ways. For example, slide 7 "CASE 4" shows the case where the level of contention is less in the global bus than it is the local bus, but the (in)coherency is greater in the global bus than the local bus. This is the kind of effect one might see with poor DB table partitioning causing more data than anticipated to be shared across the global bus. And it's precisely because it is so unintuitive that we need to do the quantitative analysis.
I intend to modify these slides to show how things scale with a variable number of users (i.e., software scalability) on a fixed hardware configuration per cluster node and present it in the upcoming GCaP class. If you have performance data for apps running on clusters of this type, I would be interested in throwing my scalability model at it.
Sunday, April 27, 2008
His comment on unit testing (or an overzealous reliance on it) seems a bit obscured here. I am certainly more inclined to concur with his other quote: "early optimization is the root of all evil," because unit testing tends to promote that bad strategy with respect to system performance. Some personal war stories describing exactly that will be delivered this week in my Guerrilla Boot Camp class.
However, do read and heed what he says about multicores and multithreaded programming. It should sound familiar. BTW, where he apparently says "Titanium", he means Itanium.
His riff on "literate programming" is a bit of a yawn for me because we had that capability at Xerox PARC, 25 years ago. Effectively, you wrote computer code (Mesa/Cedar in that case) using the same WYSIWYG editor that you used for writing standard, fully-formatted documentation. The compiler didn't care. This encouraged very readable commentary within programs. In fact, you often had to look twice to decide if you were looking at a static document or dynamic program source. As I understand it, Knuth's worthy objective is to have this capability available for other languages. The best example that I am aware of today, that comes closest to what we had at Xerox, is the Mathematica notebook. You can also use Mathematica Player to view any example Mathematica notebooks without purchasing the Mathematica product.
Wednesday, April 23, 2008
When we see a high-precision number, we perceive it as being associated with something particular (e.g., this pig). When we see a rounded number, even if it is smaller in magnitude, we perceive it as being associated with an interval or collection of numbers (e.g., the collection of pigs in the drawing). That's more or less what's going on between Sylvie and Bruno; she's thinking round numbers about the collection of pigs, whereas Bruno is focused on the particular four pigs that he can see immediately.
An interval is a range of numbers that could encompass the precise number and therefore, by definition, exceed it. So, for example, $325,425 is larger than $325,000 in absolute magnitude, but the latter could be interpreted (i.e., perceived in our mind) as representing the interval $325,000 to $326,000, which encompasses $325,425. Hence, the particular may be perceived to be smaller than the general. The Cornell researchers did not test for this effect.
Sunday, April 20, 2008
Interestingly enough, prices like $299,999 (akin to the usual sales ploy one sees in department stores and on TV) were deliberately excluded from the Cornell study because numbers ending in all 9's, although a precise number (by their definition), is considered too close to the rounded value ($300,000) to give a statistically reliable distinction.
One has to wonder, what are the implications for the way capacity planning reports are perceived? We'll discuss that point next week as part of the section on significant digits in the Guerrilla Boot Camp class.
Even if you think such statistical reports might be a bit suspect, you've gotta love their opening rubric from Lewis Carroll:
`I'm counting the Pigs in the field!' (Bruno, looking out the window)
`How many are there?' I enquired.
`About a thousand and four,' said Bruno.
`You mean "about a thousand",' Sylvie corrected him.
`There's no good saying "and four": you can't be sure about the four!'
`And you're as wrong as ever!' Bruno exclaimed triumphantly.
`It's just the four I can be sure about;
'cause they're here, grubbling under the window! It's the thousand I isn't pruffickly sure about!'
--- "Sylvie and Bruno Concluded" (1893), Ch. 5, p. 3.
Universal Scalability law, has been updated with the following diagrams,
Saturday, April 19, 2008
Call me old-fashioned, but one of things the drives me up the wall about publication on the web in general, and technical expositions in particular, is the lack of both time-stamps and citations. These two things have existed in the scientific media even before formal journal publication. For example, 17th century scientists like Newton and Hooke, wrote missives to each other and it was convention then, as it is today, to commence a letter with the date. That's how we know that Hooke was very close to coming up with the law of gravitation that is now attributed to Newton (also aided by the latter meticulously eliding all reference to Hooke after the first edition of The Principia). Could we know those things today if they had been using the Web? It's not clear. It depends. And that's the problem; lack of consistency and a lack web tools to enforce consistency.
Friday, April 11, 2008
Roberts claims, contrary to popular belief, the problem with congestion is the networks, not the TCP protocol. Rather than overhaul TCP, he says, we need to deploy flow management and selectively discard no more than one packet per TCP cycle. Flow management is the only alternative to probing into everyone's network and the only way to fairly distribute Internet capacity. The comments on his post are also worth reading because they compare his proposal with already defined protocols, such as WRED and DiffServ.
Simple registration required to download the 25 MB mp3 file. This podcast also gives you an idea of some the things we will be treating in the Guerrilla Boot Camp class on April 28-29, 2008.
Sunday, March 30, 2008
"But the underground conveyor system became clogged because staff failed to remove luggage quickly enough at the final unloading stage. So the system shut down. It's like a shopper putting too many goods on the supermarket check-out belt."
Because the baggage transfer area is tightly secured (we hope), there are no clandestine videos available from passenger cellphones. However, the following queue-theoretic simulation reveals what likely happened.
Click on the image to start the simulation.
Wednesday, March 26, 2008
Visit this group
has now been established. To join the group, just enter your email address and hit Subscribe.
Prospective authors interested in presenting a paper on PerfViz at CMG 2008, have until May 16th to submit an abstract and until June 13th to write a draft manuscript.
Tuesday, March 25, 2008
- Tick: Penryn XE, Core 2 Extreme X9000 45 nm @ 2.8 GHz
- Tock: Bloomfield, Nehalem micro-architecture 45 nm @ 3.0 GHz
Note that Sun is already shipping 8 cores × 8 threads = 64 VPUs @ 1.4 GHz in its UltraSPARC T2.
Nehalem also signals the replacement of Intel's aging frontside bus architecture by its new QuickPath chip-to-chip interconnect; the long overdue competitor to AMD’s HyperTransport bus. A 4-core Nehalem processor will have three DDR3 channels and four QPI links.
What about performance benchmarks besides those previously mentioned? I have no patience with bogus SPECxx_rate benchmarks which simply run multiple instances of a single-threaded benchmark. Customers should be demanding that vendors run the SPEC SDM to get a more reasonable assessment of scalability. The TPC-C benchmark results are perhaps a little more revealing. Here's a sample:
- A HP Proliant DL380 G5 server 3.16GHz
2 CPU × 4 cores × 1 threads/core = 8 VPU
Pulled 273,666 tpmC on Oracle Enterprise Linux running Oracle 10g RDBMS (11/09/07)
- HP ProLiant ML370G5 Intel X5460 3.16GHz
2 CPU × 4 cores × 1 threads/core = 8 VPU
Pulled 275,149 tpmC running SQL Server 2005 on Windows Server 2003 O/S (01/07/08)
- IBM eServer xSeries 460 4P
Intel Dual-Core Xeon Processor 7040 - 3.0 GHz
2 CPU × 4 cores × 2 threads/core = 16 VPU
Pulled 273,520 tpmC running DB2 on Windows Server 2003 O/S (05/01/06)
Roughly speaking, within this grouping, the 8-way Penryn TPC-C performance now matches a 16-way Xeon of 2 years ago. Note that the TPC-C Top Ten results, headed up by the HP Integrity Superdome-Itanium2/1.6GHz at 64 CPUs × 2 cores × 2 threads/core = 256 VPUs, are in the 1-4 million tpmC range.
The next step down is from 45 nm to 32 nm technology (code named Westmere), which was originally scheduled for 2013. Can't accuse Intel of not being aggressive.
Saturday, March 22, 2008
Special note added regarding J2EE/WebLogic listen threads in Chapter 7 and how they control application scalability.
You can submit an erratum online.
Wednesday, March 19, 2008
The U.S. dollar prices shown on the 2008 Training Schedule (PDF) can be converted to another currency using any of the following selection of multipliers:
Australia ..... 1.08
Brazil ........ 1.71
Canada ........ 0.99
China ......... 7.08
Euro .......... 0.63
India ......... 40.2
Japan ......... 98.23
Korea ......... 1008.4
Malaysia ...... 3.15
Switzerland ... 0.99
Taiwan ........ 30.72
Thailand ...... 31.1
U.K. .......... 0.49
If your currency is not shown here, try this online currency converter. I look forward to seeing more international students in our classes this year.
Monday, March 17, 2008
Now, we read that U.S. high-tech R&D is trending downward even further; the PC language is "narrowing our focus" . Like the continual reduction in the Fed Prime Rate, how low can you go before you have no effect? Moreover, how can such trends have been considered suicidal 25 years ago but embraced today? Japan has not gone away, and now China and India also loom as respectable tech competitors in their own right. Not being an economist, I can only think of one explanation: Wall Street. Actually, it's not exactly a new trend because it began more subtly around 15-20 years when the aforementioned corporations slowly started to divest themselves of the large-budget Edison model. Why would they do that? It's around that same time that the U.S. population at large started becoming more invested in Wall St., either directly or indirectly (e.g., retirement accounts). This trend has since been adopted in places like Europe and Australia.
The Street demands short-term gains be reported quarterly, no matter how that goal is accomplished; witness the current sub-prime banking debacle as one potential outcome of trying to meet such insane demands. But if profits in the USA are up (as they measureably are--or have been until very recently), why is there less money going into R&D? Shouldn't the percentage of profits, at least, remain constant? To answer this question, I would point to none other than "Mr. Capitalism" himself, Warren E. Buffet (see if you can guess his point before clicking on the link). Just as the deregulated financial industry has imploded (again--lest we forget the S&L debacle of the late 80's), I think Edison would be appalled at the self-destructive shift of profits out of R&D and into executive compensation packages. Sadly, it seems the Wizard of Menlo Park continues to be overshadowed by the Wizards of Wall Street. OK, end of rant.
Saturday, March 8, 2008
The above plot, showing the normalized response times (R/S) for an M/M/m queue (i.e., a single waiting line with m servers), popped up several times at Hotsos 2008. The M/M/m queue can be employed to model the performance of multiple Oracle processes. Here, the curves correspond to m = 1 (black), 2, 3, 9, 16 (blue) plotted against average server utilization.
Friday, March 7, 2008
Tuesday, March 4, 2008
Monday, February 25, 2008
"Paravirtualization has been available in the VMware Workstation product since version 6, but it also requires guest kernel changes. Linux kernel 2.6.20 and above support pv, look for CONFIG_VMI=y in kernel config. The latest Ubuntu release has this enabled by default. You also need to check the paravirtualized kernel flag in the VMware Workstation and ESX Server Advanced settings under the Options tab.
Paravirtualization in Linux has quite a few moving parts to it, and affects mostly performance metrics and timekeeping functions. The best example is %steal time, which is the time that the virtual guest wanted to run, but was not able to, usually due to contention with other VM guests. This time was previously only visible to the virtual host, but can now be seen by the guest. Paravirtualized guests do run faster, and certainly feel more responsive when using a Linux GUI like KDE on a virtual guest.
Other advances from VMware are things like "VMDesched", which is an experimental feature in VMware Tools that creates a thread that is charged all the involuntary preemption time from the physical CPU."
Thursday, February 21, 2008
Saturday, February 16, 2008
When I was at Xerox PARC, we had a term for this kind of failure mode: Error 33. Error 33 states that it is not a good idea for the success of your research project to be dependent on the possible failure of someone else's research project. This term was coined by the first Director of Xerox PARC, Dr. George Pake and the nomenclature is reminiscent of Catch 22.
Error 33 is an all too appropriate reminder that a lot of Web 2.0 technology, which is hyped as ready for prime-time, is really still in the R&D phase. It's probably only very annoying when SmugMug is off the air for several hours, but mission-critical services like banks and hospitals should approach with caution. Achieving higher reliability is only likely to come at a higher premium.
An astrophysicist recently decided to look into this question more carefully using Monte Carlo simulations and found some surprising results. He unexpectedly discovered that the common rear boarding procedure is actually the second worst procedure, since it is only slightly more efficient than boarding front-to-back! So surprised was he, that at first, he thought there was a bug in his code. Then it became apparent that there was something more subtle going on. MC sims showed that an optimal boarding method was for passengers to board 10 at a time in every other row, since loading luggage requires about two aisles of space. In this way, passengers are either stowing luggage or sitting in their seats, rather than waiting in the aisle, as they do in the other two protocols. Depending on the size of the aircraft, this boarding procedure produces a speed up of 5 to 10 times over the worst case. Of course, disembarking is still LIFO. :-)
I wonder if this result could have applications in the context of computer or network performance analysis and capacity planning?
Tuesday, February 12, 2008
Monday, February 4, 2008
The latest schedule updates can always be found on the Classes page.
Friday, February 1, 2008
- Feb, 2007 The joint announcement with IBM
- Apr, 2007 More on Penryn
- May, 2007 More on Moore
- Oct, 2007 Folsom: Not Just a Prison But A Cache
Intel is now significantly ahead of the industry with the production of 45 nm parts using the new high-K dielectric materials. The claims for the new Hafnium oxide-metal gate technology included:
- Approximatley 2x improvement in transistor density, for either
smaller chip size or increased transistor count
- Approximatley 30% reduction in transistor switching power
- Better than 20% improvement in transistor switching speed or 5x
reduction in source-drain leakage power
- Better than 10x reduction in gate oxide leakage power
Wednesday, January 30, 2008
From the performance angle, however, it may also be that the honeymoon period is now over in part because VMWare is a victim of its own success. Not only has all the hype surrounding virtualization and consolidation worn a bit thin these days, but it has also attracted the big guns---Microsoft and Oracle---into the market. And customers are not doubt becoming aware that consolidation doesn't always translate into less MIPS or more greenness. The overheads of virtualization can be very significant. The problem for us performance weenies is knowing what are those overheads in a QUANTITATIVE way. As I've stated before:
All virtualization is about illusions and although it is perfectly reasonable to perpetrate such illusions onto a user, it is entirely unreasonable to propagate those same illusions to the performance analyst.
So, here's an opportunity for VMWare to differentiate itself in the madding crowd of new VMM vendors; focus on providing more whistles and less bells. And you (Dear reader), as John Q. Customer/Analyst, should demand it (even if by proxy). If you don't make the ultimate performance issue known to management, how can they be expected to pressure the vendors?
Monday, January 28, 2008
The Nexus 7000 will deliver up to 15 Tbps of switching capacity in a single chassis, with 512 ports for 10 Gbps ethernet, and eventually it is slated to be delivered with 40 Gbps and 100 Gbps ports. Some of the claimed performance speeds-and-feeds appear rather breathtaking:
- Copy the entire Wikipedia database in 10 milliseconds.
- Copy the entire searchable Internet in 7.5 Minutes.
- Download all 90,000 Netflix movies in 38.4 seconds.
- Send a high-resolution 2 megapixel photo to everyone on earth in 28 minutes.
- Add a Web server in 9 seconds rather than 90–180 Days.
- Transmit the data in all U.S. academic research libraries (estimated at more than 2,000 TB) in 1.07 seconds.
If nothing else does it, the 3 significant digits in the last claim tells you this is marketing-speak (read: calculated using max bandwidth assumptions), so a liberal dusting of sodium chloride is recommended.
The concept of a "data center" is currently undergoing a serious transformation and it will be interesting to see how this kind of mega-switch stacks up against alternative approaches, such Google's Data Center in a Box.
Friday, January 18, 2008
Wednesday, January 16, 2008
Continuing a string of surprising announcements, Sun Microsystems today said it plans to buy the makers of MySQL open-source database software for almost $1,000,000,000 (I like to see all those zeros). Sun does have a lot of cash burning a hole in its pocket, but it also has a less than stellar track record when it comes to acquisitions.
Monday, January 14, 2008
Saturday, January 12, 2008
Please bear with us while we sort out the situation with the hotels. The Crowne Plaza now has new ownership and is undergoing a complete renovation. They are telling us they will be open again by March, but we have to decide if we need to hedge our bets on that claim. It will get sorted out shortly, but if you are planning on attending a Guerrilla class, please make sure you check the schedule page for the latest information.
Just as an aside, if you look at the Hotsos company logo at the top of their web pages, you'll see some equations or bits of equations. The first of these is the denominator of the famous Erlang-C function (A. Erlang, 1917). More on that in an upcoming blog entry.
CMG 2007. In fact, I need a vacation!!!
Anyway, I'm looking for pix (JPEGs) that people have taken over the years while attending my Guerrilla classes. In particular, one year (2003, 2005?) there were some shots taken of the entire group outside the Crowne Plaza in the afternoon sunlight. I was emailed a copy of them at the time by the person who took them, but since upgrading computers N-times, I cannot find them. :-(
Please let me know if you have any such photos.