It is proposed that an entirely new form of information processing, namely quantum computing, could be possible[1] if the states of electron spins in a given solid can be synthesised (or created), manipulated and measured at the single-quantum level. A spin-quantum dot architecture for a quantum computer, thereby indicating a variety of first generation nanostructures, is reviewed. A spin filter and spin detection mechanism [3] at the single-spin level which can be used for read-in and read-out in conventional as well as in quantum computer gates is discussed. Addressing the feasibilty of quantum communication with entangled electrons [4,5], Einstein-Podolsky-Rosen pairs are discussed.
A way of using Quantum Dots (QD) would be to look at it as a producer of electric charge and use the same electric charge as a qubit. Semiconductor QD's can serve as 3-Dimensional boxes with electrostatic potentials which confine charge quanta. Unfortunately, uncontrolled distant charged motion leads to dephasing. Also scattering reactions such as those triggered by phonon interactions cause coherence times to be relatively short for charge states.
Obtaining of Entangled Photons from Quantum Dots in a cavity for usage in Quantum Computation & Communications is steadily growing in popularity. Some progress has been done in this area [ e.g. Bensaon et. al, 2000; Stace et. al,2003; benycouf et. al., 2004; Kumar et. al., 2004]. It is also expected that there will be other such sources soon. A very demanding but also very promising area, while implementing Quantum Communication, is generating fluorescence photons by manipulating trapped atoms or ions. This is closely connected to Quantum Computation with trapped ion systems and cavity QED systems. A reference source for these techniques is, for example, "Focus On Single Photons On Demand " by Grangier et. al. 2004.
Teleportation of single qubits have already been achieved by multiple groups in laboratories using entangled photons ( for e.g., see ref. Galindo et. al below).
[1] D. Loss, D.P. DiVincenzo, Phys. Rev. A 57 (1998) 120; cond-mat/9701055.
[2] G. Burkard, H.A. Engel, D. Loss, cond-mat/0004182 (Review).
[3] P. Recher, E. V. Sukhorukov, D. Loss, cond-mat/0003089.
[4] D. Loss, E. Sukhorukov, Phys. Rev. Lett. 84, 1035 (2000).
[5] G. Burkard, D. Loss, E. Sukhorukov, to appear in Phys. Rev. B RC, cond-mat/9906071.
[6] A. Galindo and M.A. Martin-Delgado: Infomration and Computation: Classical and Quantum Aspects, 74:347-423, 2002
Wednesday, 18 February 2009
Best Bandwidth Solution(s) For Video ?
I was recently asked the following: What Are The Best Bandwidth Solution(s) For Video Conferencing & Multi-Media Applications? What bandwidth solution (T1, DS3, OCx/Sonet, etc.) would you recommend for a small, medium, and large size business ... and why .... to cover videoconferencing and multi-media applications? No other specifics to offer ...Rather I want your thoughts and recommendations for what a company should plan to (small, medium, and large).
All I could rake up from the depths of the foggyness of my mind is the following: Without knowing the specifics, it is hard to provide a precise answer, but, one can still grope around for a systematic method for calculating required bandwidth. Once you know the bandwidth requirement, then it is all about the balancing act between budget restrictions and providing the required performance. Rule of thumb, as I know it is as follows:
1. Calculate the peak external link bandwidth requirements (inter-office data transfer, video conferencing, email transfers. With attachments running in tens of megabytes, email traffic can’t be ignored these days.).
2. Real time applications being most jitter and delay sensitive - so you have to make sure that you will have enough bandwidth when they need it. The bandwidth of video depends on the mpeg profiles used (without going into specific, generally 1.5 Mbps can give you very good video on a PC (equal to VCD quality). HDTV images can take about 20 Mbps – but that is domain more reserved for IP TV service providers). Most current users of interactive video communications will be happy with the images coded and transmitted @ 512 Kbps. This includes audio and video as well as control signaling. So, one should provision at least 512 Kbps per video stream, and more the better (I would say 1.5 Mbps is the good if you are a big organization and use a large TV for video conferences)). So, multiply bandwidth for a single stream with number of parallel streams required. Now that determines the total peak real time usage.
3. There is no specific rule but wise men with experience advise to keep the peak real time within 60 to 75% of network bandwidth available leaving remaining capacity for background traffic. In a small organization of 5 people - it is easy to tell people not to download gigabyte attachments when video conferencing is going on , but in larger organization it is hard to enforce such things except with router policies (assuming they have QoS support), and you can deal with occasional unhappy users.
4. Now, once you know your bandwidth requirements, it is time to talk to the network operator how they can provide that bandwidth in the most cost effective way.
I'd say it also dedends on what kind of service and QoS you are looking at. For example, when you say "video" I suppose you mean video conferencing and not something like VoD. The Bandwidth requirements will vary according to your service requirements.
For example: For IPTV services, the image quality depends on the encoding deployed: MPEG-2 consumes approx. 3.75 Mbps, whereas MPEG-4 needs approx. 2 Mbps for the same high-quality image production. Also broadcast TV is delivered using IP Multicast which makes the bandwidth required dependent on the number of channels offered and the encoding rate. 200 channels of MPEG-2 in standard definition will take approx. 750 Mbps of bandwidth. VoD, on the other hand, is a unicast per-viewer channel. 1000 standard definition VoD users will need appro. 3.75 Mbps.
The QoS requirements for video conferencing using H.323 ( SIP could be different again) can be planned on the "Rule of 75" as follows: Calculate the minimum bandwidth required b each of your application( e.g., video, voice, data). The total of this badnwidth is the minimum requirement for any given link and it should consume NO MORE than 75% of the total available bandwidth on the link. The 75% rule makes allowances for bandwidth required for over head traffic, such as routing, Layer 2 keepalives and other applications, such as, email, HTTP etc.
So, Capacity planning for H.323, should look like something as follows:
Video data + 20% = bandwidth required.
Example Video data rate: Bandwidth Required: 512 kbps = 614 kbps ; 1.5Mbps = 1.8 Mbps ...
For issues such as number of concurent users and more stuff on video conferencing you can perhaps consider looking into Cisco's solutions offered and also TANDBERG boxes.
All I could rake up from the depths of the foggyness of my mind is the following: Without knowing the specifics, it is hard to provide a precise answer, but, one can still grope around for a systematic method for calculating required bandwidth. Once you know the bandwidth requirement, then it is all about the balancing act between budget restrictions and providing the required performance. Rule of thumb, as I know it is as follows:
1. Calculate the peak external link bandwidth requirements (inter-office data transfer, video conferencing, email transfers. With attachments running in tens of megabytes, email traffic can’t be ignored these days.).
2. Real time applications being most jitter and delay sensitive - so you have to make sure that you will have enough bandwidth when they need it. The bandwidth of video depends on the mpeg profiles used (without going into specific, generally 1.5 Mbps can give you very good video on a PC (equal to VCD quality). HDTV images can take about 20 Mbps – but that is domain more reserved for IP TV service providers). Most current users of interactive video communications will be happy with the images coded and transmitted @ 512 Kbps. This includes audio and video as well as control signaling. So, one should provision at least 512 Kbps per video stream, and more the better (I would say 1.5 Mbps is the good if you are a big organization and use a large TV for video conferences)). So, multiply bandwidth for a single stream with number of parallel streams required. Now that determines the total peak real time usage.
3. There is no specific rule but wise men with experience advise to keep the peak real time within 60 to 75% of network bandwidth available leaving remaining capacity for background traffic. In a small organization of 5 people - it is easy to tell people not to download gigabyte attachments when video conferencing is going on , but in larger organization it is hard to enforce such things except with router policies (assuming they have QoS support), and you can deal with occasional unhappy users.
4. Now, once you know your bandwidth requirements, it is time to talk to the network operator how they can provide that bandwidth in the most cost effective way.
I'd say it also dedends on what kind of service and QoS you are looking at. For example, when you say "video" I suppose you mean video conferencing and not something like VoD. The Bandwidth requirements will vary according to your service requirements.
For example: For IPTV services, the image quality depends on the encoding deployed: MPEG-2 consumes approx. 3.75 Mbps, whereas MPEG-4 needs approx. 2 Mbps for the same high-quality image production. Also broadcast TV is delivered using IP Multicast which makes the bandwidth required dependent on the number of channels offered and the encoding rate. 200 channels of MPEG-2 in standard definition will take approx. 750 Mbps of bandwidth. VoD, on the other hand, is a unicast per-viewer channel. 1000 standard definition VoD users will need appro. 3.75 Mbps.
The QoS requirements for video conferencing using H.323 ( SIP could be different again) can be planned on the "Rule of 75" as follows: Calculate the minimum bandwidth required b each of your application( e.g., video, voice, data). The total of this badnwidth is the minimum requirement for any given link and it should consume NO MORE than 75% of the total available bandwidth on the link. The 75% rule makes allowances for bandwidth required for over head traffic, such as routing, Layer 2 keepalives and other applications, such as, email, HTTP etc.
So, Capacity planning for H.323, should look like something as follows:
Video data + 20% = bandwidth required.
Example Video data rate: Bandwidth Required: 512 kbps = 614 kbps ; 1.5Mbps = 1.8 Mbps ...
For issues such as number of concurent users and more stuff on video conferencing you can perhaps consider looking into Cisco's solutions offered and also TANDBERG boxes.
Tuesday, 10 February 2009
Unified Computing and "Network is the Computer" ?
Cisco's CTO, Ms. Padmasree Warrior wrote in her blog that “Cisco believes that the network can be a focal point for innovation, helping us enter new and adjacent markets if and when there is the right combination of value proposition and receptive audience.” Ms. Warrior also said “One point I will reiterate is that we see the need for open and consistent standards at many different layers in our developments—from management to protocols to data link definitions to physical layer and power/cooling standards. There is always a balance between innovation and standards, but I feel that in some areas, like Unified Computing, we can can achieve both.”
Personally, for whatever it is worth, I cannot agree more with Ms. Warrior. I am a Cisco fan, having worked extensively with Cisco’s technology for the last 9 odd years. I have reached a point where almost anything Cisco does or touches, I believe they will get the gold. A dissappointment in recent years has been their WAN acceleration areas and persistent CPU issues on some of their other series. But those are small fractions of Cisco’s core market focus. However, if they do go for stuff like data centre virtualisation and unified computing, then they will probably feel compelled to sort the WAFS/WAAS area out as well. May not be easy for Cisco to achieve all of the above with a well advertised hiring freeze on. Could be a delicate balancing issue. Good part on Cisco’s side is they probably have more time than “usual” due to less demand because of spending cuts at most of their customers.
I also believe that healthy competition in technical innovation is good for customers and future of technology. I also do not see any viable reason why people and some competitors of Cisco should receive this move of their’s as aggressively as some of them seem to have.
I tend to agree with all of our colleagues who believe that stuff like combination of propreitory and open source technologies, Cloud computing and network based computing can and will be part of a futuristic technological innovative process. Besides keeping in mind the issues of freedom and versatility that these tend to offer, the technical implementation, from a user’s point of view, so far, has been reasonably simple. An article from IBM about Amazon’s cloud computing services touches on the same.
What has been tedious for amateur technical enthusiasts like myself is the restrictions put on by propreitory licensing. I for exmaple, have tried for awhile to break into applied Quantum Cryptography across networks as I do believe that sooner or later the breakthrough has to come in Quantum Communications and the almost infinite possibilities that show up with it. The part where it may face a hitch is whether the reknowned vendors or Internet communication equipment makers (with the big money…) feel compelled to invest enough behind R&D of the same. The current limitation is the tarnsformation of the theory into efficient practice. However, implmentation of even a simple key exchange script has been challenging simply because I use Cisco and either my scripts don’t work or Cisco’s IOS coding is not compatible with mine. There is no way for me to know as I have no access to their IOS codes. Hence I feel trumped. The point being such restrictions put people without access to extensive R&D facilities at a loss. Of course since that is not a big loss for technology and business, it is unimportant.
Besides further development related to VoIP, VoD, IPv6 etc., I strongly believe that a far more secure internet is necessary. Currently, the financial world is already active in migrating to dedicated fibres due to security risk (SWIFT network). What we very definitely need in future is (almost) unbreakable security and total privacy. I believe what we need to seriously invest into and develop is the applied part of quantum information theory[1] and cryptography alongwith the related disciplines of steganography, traffic security and cryptosystems as applied towards discreet communication. In quantum computing, the laws of physics protect the information using the properties of quantum me- chanics. Open-air quantum key distribution with single photon source (SPS) has been demonstrated in experimental conditions[2][3][4]. In the area of Quantum Cryptography, in particular, it has been shown that there are intrinsic properties in Quantum Mechanics that will enable a Quantum Computer to produce results not possible with a classical computer[5][6][7]. In fututre we need to investigate the concept of development of Internet technologies based on the quantum principles[8] of cryptography, secret sharing and teleportation.
In particular, the possibility of quantum optics giving birth to a new generation of communication protocols over internet has to be seriously looked into and researched. When we do this we shall also need to address the complications arising from the atomic level structure of a quantum computing system such as: decoherence, entanglement, quantum teleportation[9], unitary transformations, and reversible universal gate structures.
Personally, for whatever it is worth, I cannot agree more with Ms. Warrior. I am a Cisco fan, having worked extensively with Cisco’s technology for the last 9 odd years. I have reached a point where almost anything Cisco does or touches, I believe they will get the gold. A dissappointment in recent years has been their WAN acceleration areas and persistent CPU issues on some of their other series. But those are small fractions of Cisco’s core market focus. However, if they do go for stuff like data centre virtualisation and unified computing, then they will probably feel compelled to sort the WAFS/WAAS area out as well. May not be easy for Cisco to achieve all of the above with a well advertised hiring freeze on. Could be a delicate balancing issue. Good part on Cisco’s side is they probably have more time than “usual” due to less demand because of spending cuts at most of their customers.
I also believe that healthy competition in technical innovation is good for customers and future of technology. I also do not see any viable reason why people and some competitors of Cisco should receive this move of their’s as aggressively as some of them seem to have.
I tend to agree with all of our colleagues who believe that stuff like combination of propreitory and open source technologies, Cloud computing and network based computing can and will be part of a futuristic technological innovative process. Besides keeping in mind the issues of freedom and versatility that these tend to offer, the technical implementation, from a user’s point of view, so far, has been reasonably simple. An article from IBM about Amazon’s cloud computing services touches on the same.
What has been tedious for amateur technical enthusiasts like myself is the restrictions put on by propreitory licensing. I for exmaple, have tried for awhile to break into applied Quantum Cryptography across networks as I do believe that sooner or later the breakthrough has to come in Quantum Communications and the almost infinite possibilities that show up with it. The part where it may face a hitch is whether the reknowned vendors or Internet communication equipment makers (with the big money…) feel compelled to invest enough behind R&D of the same. The current limitation is the tarnsformation of the theory into efficient practice. However, implmentation of even a simple key exchange script has been challenging simply because I use Cisco and either my scripts don’t work or Cisco’s IOS coding is not compatible with mine. There is no way for me to know as I have no access to their IOS codes. Hence I feel trumped. The point being such restrictions put people without access to extensive R&D facilities at a loss. Of course since that is not a big loss for technology and business, it is unimportant.
Besides further development related to VoIP, VoD, IPv6 etc., I strongly believe that a far more secure internet is necessary. Currently, the financial world is already active in migrating to dedicated fibres due to security risk (SWIFT network). What we very definitely need in future is (almost) unbreakable security and total privacy. I believe what we need to seriously invest into and develop is the applied part of quantum information theory[1] and cryptography alongwith the related disciplines of steganography, traffic security and cryptosystems as applied towards discreet communication. In quantum computing, the laws of physics protect the information using the properties of quantum me- chanics. Open-air quantum key distribution with single photon source (SPS) has been demonstrated in experimental conditions[2][3][4]. In the area of Quantum Cryptography, in particular, it has been shown that there are intrinsic properties in Quantum Mechanics that will enable a Quantum Computer to produce results not possible with a classical computer[5][6][7]. In fututre we need to investigate the concept of development of Internet technologies based on the quantum principles[8] of cryptography, secret sharing and teleportation.
In particular, the possibility of quantum optics giving birth to a new generation of communication protocols over internet has to be seriously looked into and researched. When we do this we shall also need to address the complications arising from the atomic level structure of a quantum computing system such as: decoherence, entanglement, quantum teleportation[9], unitary transformations, and reversible universal gate structures.