标签: modeling

Most electronic, mechanical, and thermal engineers think of how to keep the temperature of their IC or printed circuit board below some maximum allowable value. Others are more worried about the overall enclosure, which can be range from a self-contained package such as a DVR to a standard rack of boards and power supplies.   Basic techniques for getting heat from an IC, board, or enclosure involve one or more of heat sinks, heat spreaders (PC-board copper), head pipes, cold plates, and fans; it can sometimes move up to more-active cooling approaches including air conditioning or embedded pipes with liquid flow. That's all well and good, but obviously not good enough for the megawatts of a "hyperscale" data center. (If you are not sure what a hyperscale data center is, there's a good explanation here ). While there is no apparent formal standard on the minimum power dissipation to be considered hyperscale, you can be sure it's in the hundreds of kilowatt to megawatt range.   But where does all that heat go? Where is the "away" to which the heat is sent? If you’re cooling a large data center, that "away" to hard to get to, and doesn't necessarily want to take all that heat you are dissipating.   A recent market study from a BSRIA offered some insight in the hyperscale data-center cooling options and trends. I saw a story on the report in the November issue of Cabling Installation Maintenance , a publication which gives great real-world perspective into the nasty details of actually running all those network cables, building codes, cabling standards, and more. (After looking through this magazine you'll never casually say, it’s "no big deal, it’s just an RJ-45 connector" again.)   BSRIA summarized their report and used a four-quadrant graph (below) of techniques versus data-center temperatures to clarify what is feasible and what is coming on strong. Among the options are reducing dissipation via variable-speed drives and modular DC supplies, cooling techniques from liquid cooling to adiabatic evaporative cooling, or allowing a rise in server-inlet temperature. The graph also shows the growth potential versus investment level required for each approach; apparently, adiabatic/evaporative cooling is the "rising star."   Cooling approaches for hyperscale data centers encompass basic dissipation reduction, liquid cooling, and adiabatic/evaporative cooling, according to this analysis from BSRIA Ltd, with the latter a "rising star."   When you are worried about cooling your corner of a PC board, it's easy to forget that it's not enough to succeed in that goal; you have to think of the next person who will have to deal with the heat which you so nicely spirited away. That's why I am often wary of PC-board heat spreaders, unless the design has been thermally modeled "across the board", so to speak: they move the heat from your IC to the next one, and so make their thermal headache more difficult.   Although I know I won't be involved with design of such hyperscale cooling, I need to learn more about thermal principles, including adiabatic/evaporative cooling. It still hurts that a very long time ago, when I was told to take an engineering course on "thermal physics" to learn basics, I was misled. It turned out the course was about the personal thermal lives of atoms and molecules, and had nothing at all to do with "thermal" as engineers need to know it: heat, heat flow, thermal modeling, temperature rise, cooling techniques, and more. In contrast, "thermal physics" is what Einstein used in one his is five 1905 papers, " Über die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen " ("On the Motion of Small Particles Suspended in a Stationary Liquid, as Required by the Molecular Kinetic Theory of Heat"), but hey, he wasn't worried about cooling a hot component!

Recently J P Morgan announced the use of an "FPGA Supercomputer" to cut risk calculation times from 8 hours to under 4 minutes. Calxeda announced low energy servers based on traditional embedded architectures, targeted for data centers. Tablet computers are already replacing desktops and laptops for many applications. These trends have one thing in common - traditional embedded systems are slowly and steadily taking over not only the traditional desktop and server but also entering the domain of the mainframe and super computer ! These seemingly different manifestations of embedded systems can potentially share the same underlying design paradigm. When embedded designers start creating designs for smart devices, high performance devices and data center blades using the same common set of components, there will surely be disruption in the market place for all kinds of computing devices. How will modeling languages and design platforms look like that will power these design teams ? Fortunately, the basic elements for such design platforms are already available today, to a different level of completeness from various existing ESL/EDA vendors. I would say that the design platform for application driven embedded system level design should necessarily support the following features: A) Languages Support a family of simple and elegant architectural modeling and specification languages for + Hardware system modeling + Processor architecture modeling + Software behavior modeling + Unified system (hardware, software, communication, parallelization) configuration modeling + Unified system deployment modeling B) Functionality + Processor and System Synthesis + Automated Synthesis of Verification suite + Virtual Prototyping (for performance analysis and software development, concurrent design) + Infrastructure Virtualization (Realize parts of the system on a processor or FPGA Cloud)   + Parallelization Control Synthesis (Message exchange between processors/tasks/nodes) + SDK Synthesis + Deployment Plan Synthesis ( System level partitioning and scheduling ) Each of these distinct areas would include several tools to automate development and testing activities. Some of these would be used during specification time, others at execution time. Go ahead and start creating specs for your own organizations large system modeling platform.  

Embedded Security is increasingly becoming an important area of concern. Embedded security encompasses components, sub-systems and networks of systems. Modern embedded systems incorporate several components including many that are active. Software and hardware from different sources is integrated into a single system often by teams that may not fully understand or have little control over the functioning of the individual components. This poses a major threat and area of concern for system architects. System level design offers a significant advantage in terms of allowing a system architect to incorporate a security framework that can impose system and inter-system level security constraints. A system level security framework will need to be architectural model driven, "synthesizable" to the security model of the hardware or software level implementation and easily integrate into the inter-system level security standards already in place. A component / thread requesting a service on a modern embedded system may have little knowledge of where that request will be transported and processed. Some constraints are required on where it CAN be sent by the system - sort of a Security Impedance Policy. A component processing a request may have little knowledge about the origin of the request. Some constraints are required on what credentials and rights need to be presented that will allow the request to be processed, a Security Access Policy There needs to be an association between principals - machines or humans, and their run-time credentials ( id, authentication level, authentication status assigned roles) - Identity and Authentication Policy and their privileges in a particular sub-system - Domain Access Policy Finally in really complex systems there needs to be a run-time component that can return a dynamic "domain access policy" for individual components, preferably at system startup. Those principals, credentials, privileges, identity and authentication, security  access policy model and domain access policy are part of enterprise class systems already. The security impedance policy may be unique to embedded systems, even though it will increasingly become important in enterprise and Internet grade systems (example content at what level of trust must be returned by a google search ?). Early unix systems did an excellent job of simplifying the security model for the then security requirements into all of 9 bits (rwxrwxrwx). Today every system architect faces the challenge of creating a simple model of security (The Security Compuational Model) for their architecture. Trust (!) me - the null security model may just be the single reason for your entire system to become obsolete and insecure leading to complete product failure in the very near future.  

Please try to remember my blog on the three puzzles to ponder . Someone wrote me an email with a question regarding this blog to which I'm sure I know the answer, but...   As you may recall, the third puzzle was concerned with Gravity, which – of course – explains the title of this blog. What do you mean "What does gravity have to do with dogs and the groins of strangers?" What do they teach the young people of today? Who amongst us could forge the immortal words of the great American philosopher Dave Barry, who famously said: "Magnetism is one of the six fundamental forces of the universe, with the other five being gravity, Duct tape, whining, remote control, and the force that pulls dogs toward the groins of strangers."   But we digress... the problem I originally posed was to assume that the circumference of the world is exactly 24,000 miles at the equator. We know that the earth spins round once every 24 hours, so at the equator this is 24,000 miles in 24 hours, which equates to a rotational speed at the surface of 1,000 miles per hour (no wonder my hair always looks so mussed up).   Anyway, the question I posed was as follows: "If you are standing at the equator, the spin of the world is sort of trying to throw you off (if you see what I mean). So, assuming that someone has a mass of exactly 100kg if we were to weigh him or her at the North or South Pole, what would the reading on the weighing scale be if our subject was standing on it at the equator?"     Over time we added all sorts of qualifiers, such as the weighing scale being calibrated in such a way that being moved to the equator did not affect it, and so on and so forth.   The thing is that someone who shall remain nameless to protect the innocent – let's call him Simon (you get 10 extra points if you correctly identify the animated film that prompted me to use the name Simon) – emailed me with a query, that spawned a flurry of messages, which evolved into a telephone converzation, whose conclusion left both of us scratching our heads...   Simon: There is of course the difference between mass (an inherent property of matter, independent of gravitational field, velocity, or acceleration) and weight (which is the force felt by the Earth pushing up on your feet, dependent on your mass and the Earth's gravitational field and any vertical acceleration). The mass doesn't change. The gravitational field of the earth doesn't change. Since there is no vertical acceleration, the measured weight will be the same. That's my theory, anyway.   Max: But the Earth is spinning, so the guy will weigh less at the equator than at one of the poles because the spinning earth causes centrifugal (or centripetal?) force...   Simon: Suppose you were standing at the equator when the speed of the Earth's rotation suddenly doubled – you would fall over backwards because the acceleration would be in the horizontal direction – this is elementary stuff they teach at high school.   Max: I must have been off school that day. So here's another example, suppose I have a rock on the end of a piece of string and I'm spinning it round and round really fast in a vertical plane such that it passes my feet and then goes above my head...     Max: So now let's assume that the string breaks (or is cut by a razor) at the point when the rock is directly above my head. In this case the rock has a mix of inertia in the vertical direction and motion in the horizontal direction, so it will fly up and across, sort of thing...     Simon: Of course it won't. At the point that the rock is directly overhead, the only motion it has is horizontal, so it will take off in a horizontal direction:     Max: You are making my head hurt. Let's return to my original problem, but let's assume that there is no atmosphere. Suppose I'm standing on the equator of an airless object like the moon that is spinning really, really quickly. Are you saying that I won't weigh less at the equator than I do at one of the poles?   Simon: That's right.   Max: Suppose that it's spinning incredibly quickly – like 100,000 times an hour, surely it would fling me off into space.   Simon: No it wouldn't because all you would have would be horizontal motion, but no vertical acceleration.   Max: Arrrgggh! Look, imagine that you're looking down on the equivalent of the North Pole on the moon and you see me standing in a spacesuit on the equator. Let's assume that, from your point of view, I'm standing on top of the moon, which is spinning at some rate in a clockwise direction:     Max: If the moon suddenly disappeared at this point in time, I agree that I would carry on traveling in a horizontal direction off to the right. Now, even if the moon doesn't disappear, inertia will cause my body to want to continue travelling in a horizontal direction. However the gravity of the moon will pull me down, but due to my horizontal motion I will appear to be lighter – that is, although I will have the same mass I will appear to weigh less. If I gradually kept on increasing the rotational speed of the moon, there would come a time when my tendency to keep travelling horizontally (my inertia coupled with my horizontal velocity) would be more than the gravity of the moon and my feet would leave the ground.     Simon: No they wouldn't.   Max: Yes they would.   Simon: No they wouldn't.   Max: Yes they would.   ... This is pretty much where we left things ... with me obviously winning the argument (grin) ...   Actually, now I come to think about it, I could use that Physics modeling program Phun to model the one about the string with the rock (. Well, I could if I had the time, but as usual I am up to my ears in alligators fighting fires without a paddle ... maybe you could use Phun to test this out and report back to the rest of us...

Please try to remember my blog on the three puzzles to ponder. Someone wrote me an email with a question regarding this blog to which I'm sure I know the answer, but...   As you may recall, the third puzzle was concerned with Gravity, which – of course – explains the title of this blog. What do you mean "What does gravity have to do with dogs and the groins of strangers?" What do they teach the young people of today? Who amongst us could forge the immortal words of the great American philosopher Dave Barry, who famously said: "Magnetism is one of the six fundamental forces of the universe, with the other five being gravity, Duct tape, whining, remote control, and the force that pulls dogs toward the groins of strangers."   But we digress... the problem I originally posed was to assume that the circumference of the world is exactly 24,000 miles at the equator. We know that the earth spins round once every 24 hours, so at the equator this is 24,000 miles in 24 hours, which equates to a rotational speed at the surface of 1,000 miles per hour (no wonder my hair always looks so mussed up).   Anyway, the question I posed was as follows: "If you are standing at the equator, the spin of the world is sort of trying to throw you off (if you see what I mean). So, assuming that someone has a mass of exactly 100kg if we were to weigh him or her at the North or South Pole, what would the reading on the weighing scale be if our subject was standing on it at the equator?"     Over time we added all sorts of qualifiers, such as the weighing scale being calibrated in such a way that being moved to the equator did not affect it, and so on and so forth.   The thing is that someone who shall remain nameless to protect the innocent – let's call him Simon (you get 10 extra points if you correctly identify the animated film that prompted me to use the name Simon) – emailed me with a query, that spawned a flurry of messages, which evolved into a telephone converzation, whose conclusion left both of us scratching our heads...   Simon: There is of course the difference between mass (an inherent property of matter, independent of gravitational field, velocity, or acceleration) and weight (which is the force felt by the Earth pushing up on your feet, dependent on your mass and the Earth's gravitational field and any vertical acceleration). The mass doesn't change. The gravitational field of the earth doesn't change. Since there is no vertical acceleration, the measured weight will be the same. That's my theory, anyway.   Max: But the Earth is spinning, so the guy will weigh less at the equator than at one of the poles because the spinning earth causes centrifugal (or centripetal?) force...   Simon: Suppose you were standing at the equator when the speed of the Earth's rotation suddenly doubled – you would fall over backwards because the acceleration would be in the horizontal direction – this is elementary stuff they teach at high school.   Max: I must have been off school that day. So here's another example, suppose I have a rock on the end of a piece of string and I'm spinning it round and round really fast in a vertical plane such that it passes my feet and then goes above my head...     Max: So now let's assume that the string breaks (or is cut by a razor) at the point when the rock is directly above my head. In this case the rock has a mix of inertia in the vertical direction and motion in the horizontal direction, so it will fly up and across, sort of thing...     Simon: Of course it won't. At the point that the rock is directly overhead, the only motion it has is horizontal, so it will take off in a horizontal direction:     Max: You are making my head hurt. Let's return to my original problem, but let's assume that there is no atmosphere. Suppose I'm standing on the equator of an airless object like the moon that is spinning really, really quickly. Are you saying that I won't weigh less at the equator than I do at one of the poles?   Simon: That's right.   Max: Suppose that it's spinning incredibly quickly – like 100,000 times an hour, surely it would fling me off into space.   Simon: No it wouldn't because all you would have would be horizontal motion, but no vertical acceleration.   Max: Arrrgggh! Look, imagine that you're looking down on the equivalent of the North Pole on the moon and you see me standing in a spacesuit on the equator. Let's assume that, from your point of view, I'm standing on top of the moon, which is spinning at some rate in a clockwise direction:     Max: If the moon suddenly disappeared at this point in time, I agree that I would carry on traveling in a horizontal direction off to the right. Now, even if the moon doesn't disappear, inertia will cause my body to want to continue travelling in a horizontal direction. However the gravity of the moon will pull me down, but due to my horizontal motion I will appear to be lighter – that is, although I will have the same mass I will appear to weigh less. If I gradually kept on increasing the rotational speed of the moon, there would come a time when my tendency to keep travelling horizontally (my inertia coupled with my horizontal velocity) would be more than the gravity of the moon and my feet would leave the ground.     Simon: No they wouldn't.   Max: Yes they would.   Simon: No they wouldn't.   Max: Yes they would.   ... This is pretty much where we left things ... with me obviously winning the argument (grin) ...   Actually, now I come to think about it, I could use that Physics modeling program Phun to model the one about the string with the rock (. Well, I could if I had the time, but as usual I am up to my ears in alligators fighting fires without a paddle ... maybe you could use Phun to test this out and report back to the rest of us...