APOLLO:Multiplexed Lunar Laser Ranging T.W.Murphy,Jr.,E.G.Adelberger,J.D.Strasburg,&C.W.Stubbs Dept.of Physics,University of Washington,Seattle,WA98195-1560
Abstract
The Apache Point Observatory Lunar Laser-ranging Operation(APOLLO)is a next-generation lunar laser ranging(LLR)campaign aimed at order-of-magnitude improvements
in tests of gravitational physics via millimeter range precision.We will employ the3.5m
telescope at the Apache Point Observatory(APO),located in southern New Mexico at an
altitude of2800m.As a result of the large aperture size and excellent seeing conditions
at the site,APOLLO expects to detect2–10lunar return photons per pulse.Relative
background immunity permits operation in daylight and at full moon,resulting in better
sampling of the lunar orbit.
We will use avalanche photodiode(APD)arrays as the detector for APOLLO,allow-ing multiple photons within a single return pulse to be individually time-tagged with high
precision.Immediate advantages are shot-by-shot range pro?les,as well as guaranteed cal-
ibration for each shot.In conjunction with a high time-precision start photodiode,one
quickly builds an accurate representation of the convolved laser-detector-electronics tempo-
ral response.We describe here the hierarchical,multiplexed APOLLO timing system and
its implementation using commercial electronics and a programmable logic controller.We
also discuss the various calibration procedures geared toward millimeter precision.There
is no practical limit to the number of channels employed in our timing scheme,allowing an
upgrade path to larger APD array formats(10×10or larger)in the future.
1Introduction
Lunar laser ranging(LLR)has provided a laboratory for tests of fundamental physics for over three decades[1,2].Centimeter-precision range information provides measurements at the part in 1013level by comparison to solar-system-scale phenomena.As we build the lunar-range database, some measurements become increasingly sensitive with the passage of time even without im-provements in the range precision.For example,sensitivity to the time variation of Newton’s gravitational constant,G,improves as the square of the observation time—now constrained at the impressive level of less than a part per trillion change per year[3].
While LLR may still deliver improved tests of physics in the years to come,the poten-tial for order-of-magnitude improvements is best enabled by a new campaign intent on making order-of-magnitude improvements to the range precision itself.Current systems deliver range normal-points with typical uncertainties around2cm[4].The greatest impediment to substan-tial improvement in range precision is photon number.Because the strength of the return signal from a passive re?ector varies as the inverse fourth-power of distance,laser ranging to the moon is at the very limit of detectability.For example,the McDonald Laser Ranging Station(MLRS, [5])at its very best detects just over one photon per minute(600laser shots),placing about15 photons into one normal point.From a purely statistical standpoint,order-of-magnitude range improvements would necessitate～1500photons per normal point—well beyond the capabilities of current LLR stations.
By establishing lunar range capability on the3.5m telescope at the Apache Point Observatory, we anticipate moving into the multiple-photon-per-pulse regime[6].At a20Hz pulse repetition rate,the requisite photon number is accumulated in a matter of minutes.However,the high photon return rate presents instrumental challenges if one is to measure the lunar distance in an unbiased way.In particular,a signal consisting of a few photons is not large enough to reliably de?ne a pulse shape on which one could perform a meaningful centroid or leading-edge measurement.Nor is this signal strength suitable for single-photon detectors such as avalanche photodiodes(APDs),which ignore all but the?rst detected photon.We will describe here the technique adopted for APOLLO to perform an unbiased measurement on a pulse containing a small number of photons.This paper will largely concentrate on the timing scheme employed to accommodate an arbitrary number of single-photon detector channels,which for APOLLO assumes the form of an APD array.
APOLLO expects to deliver range precision in the neighborhood of one millimeter.This high level of precision will undoubtedly expose many model de?ciencies,almost all of which will be associated with earth deformations and atmospheric in?uence.Even before the model has time to catch up to APOLLO’s precision,we will be able to extract gravitational physics measurements out of the dataset,since these appear at very well-de?ned frequencies,unlikely to be mimicked by aperiodic backgrounds or by periodic in?uences at di?erent frequencies.Within the course of several years,APOLLO will produce order-of-magnitude improvements in fundamental physics parameters.Speci?cally,APOLLO will provide the very best tests of the following aspects of gravity,to the precisions indicated:
?Test of the strong equivalence principle to3×10?5
?Test of the weak equivalence principle with?a/a≈10?14
?Measurement of˙G/G to10?13yr?1
?Test of the1/r2law at the lunar distance scale to～10?12
?Measurement of geodetic(de Sitter-Fokker)precession to～3×10?4
The test of the strong equivalence principle tests how gravitational energy itself gravitates—a crucial nonlinearity predicted by general relativity[7,8,9].The weak equivalence principle probes compositional di?erences in the accelerations of two bodies,and will exceed laboratory tests of the same[10].Characterization of˙G addresses the evolution of fundamental coupling constants against the backdrop of the expanding universe[11,e.g.].The test of the inverse square law searches for new long-range forces,and may be able to subject certain brane-world cosmological scenarios to critical tests[12,13].Measurement of relativistic geodetic precession provides a measure of the curvature of local spacetime,and the e?ect this has on the orientation of coordinate systems[14].
In addition to the aformentioned gravitational measurements,APOLLO will deliver order-of-magnitude gains in lunar science,geodesy,coordinate determinations,etc.Especially improved will be lunar science,because the high photon return rate will enable range measurements to the complete set of lunar re?ectors during each observation.Thus the lunar orientation and deformation will be very precisely determined.
2APOLLO Photon Rate
The high photon rate expected from APOLLO is the key to achieving substantial gains in pre-cision.The principal factors enabling APOLLO to accomplish its millimeter goal are the large telescope aperture and the excellent atmospheric seeing experienced at the site.The median net image quality of the APO3.5m telescope for long exposures is1.05arcseconds.A crude scaling of APOLLO to OCA(Observatoire de la C?o te d’Azur:1.5m,2.5arcsecond image quality [15])and MLRS(0.78m,4arcsecond image quality[5])indicates a signal gain of35and300, respectively.A gain of this magnitude(roughly102)is necessary for a single order-of-magnitude improvement in random range uncertainty.
A bottom-up calculation of the expected photon return rate based on the familiar link ef-?ciency equation is also illuminating.For APOLLO’s parameters,we calculate the following number of return photons per laser pulse:
N det=5.4× E
115mJ
η
0.4
2 f
0.25
Q
0.3
n
100
1asec
Φ
2 10asec
φ
2
r
r
4
photons.
Here,E is the pulse energy,ηis the one-way optical e?ciency,f is the receiver throughput (dominated by narrow-band?lter),and Q is the detector e?ciency.We assume one arcsecond seeing,pointing toward the smaller of the Apollo arrays(100array elements),10arcsecond divergence out of the corner cubes,and the moon at its mean distance.We have been intentionally pessimistic with regard to optical e?ciencies,but nonetheless calculate a detection rate of5 photons per pulse.This return rate exceeds that routinely experienced at OCA and MLRS by a factor of roughly1000—at odds with the simpler scaling in the previous paragraph.
Although the link e?ciency equation as applied to current LLR stations yields photon rates roughly an order-of-magnitude too optimistic for current LLR stations,this is not the case when applied to the LLR system that once operated on the2.7m telescope at McDonald Observatory. At roughly0.8photons per pulse,this system had a high enough photon return rate to enable real-time optimization of the system[16].APOLLO will likewise operate in this regime.Even a factor of100degradation in e?ciency(due to pointing,focus,etc.)will produce one photon per second—enough to attempt real-time adjustment of pointing,focus,transmit/receive misalignment,and beam collimation.All of these adjustments are actuated in APOLLO,allowing an automated optimization routine to?nd the parameters that yield peak performance.
Another indication that this expectation may be reasonable comes from the best historical performance of the MLRS and OCA LLR stations.Without the ability to?ne-tune performance in a closed-loop,algorithmic manner,the idealized system capabilities are perhaps best charac-terized by the lucky occasions when the systems happened to be tuned to optimum performance. MLRS,for instance,reports receiving120photons in41minutes on the Apollo15re?ector[17]. Considering the time spent steering the telescope,one might estimate a peak instantaneous rate of one photon every10seconds,or0.01photons per pulse.For OCA,a return rate of0.1pho-tons per pulse was obtained on the Apollo15re?ector[18].Blindly scaling these rates by the factors computed above based on aperture and seeing alone,one predicts a peak photon rate for APOLLO of about3photons per pulse.
So far,the issue of background noise has been left out of the discussion,but here too APOLLO enters a new regime—owing largely to the very small accepted?eld of view of1.2arcsec.Because of this,APOLLO will be able to operate in full moon or daylight conditions without signi?cant noise contamination.We predict about0.2photons per pulse per100ns gate,on the full moon.
Figure1:The4×4APD array APOLLO will use,courtesy of Lincoln Labs.The array elements are30μm in diameter,separated by100μm.Spatial information is preserved in the pixelated grid,and may be used to perform real-time guiding on the return signal.
Within the～1ns spread of lunar photon returns,our signal-to-noise ratio is～500with a signal return as low as one photon per pulse.
3Array Detector
Traditionally,laser ranging(e.g.,to satellites)has operated in either the single-photon mode—relying on the assumption that“double”photons are rare—or in large-signal mode,whereby the pulse shape is represented directly in the detector response.For our intermediate photon rate,neither is particularly suitable.Ideally,we would like to time-tag each detected photon individually.A single detector is incapable of this,because the detector response time is longer than the duration of the photon bundle.A multiplexed scheme is called for.One technique would string many single-element detectors along a chain of beam-splitters,such that each detector statistically receives some fraction of the incident light.Another technique is to use an array detector,spreading the return light spatially across the array such that any given element is again statistically likely to receive some fraction of the input light.
In collaboration with MIT Lincoln Laboratories,we have access to the APD array technology developed there.By fabricating APD structures into a monolithic silicon substrate,they have produced4×4and32×32array formats(see Figure1).The elements have active areas of20, 30,or40microns,placed on a square array pattern with100μm spacing.Details about the detector arrays may be found elsewhere[19].
The APD array that will initially be used in APOLLO is a4×4format with30μm elements (Figure1).Because we want each of the elements to operate in single-photon mode(or an approximation to this),we want the likelihood of photon detection per element to be<0.25, or about4photons across the array per pulse.At this upper limit,25%of the detections will statistically be“double”detections,and a substantial bias correction must be applied—less for weaker returns.It has yet to be determined how high this upper limit should be for recovery of millimeter net accuracy in locating the geometrical center of the retrore?ector array palette.The “comfort-zone”for an array of this size is around1–3photons per pulse,right where APOLLO expects to operate.
4Multi-photon Calibration
The technique of measuring the time of departure of the outgoing laser pulse through the use of a corner-cube or other re?ector in the telescope exit aperture is here called calibration .Serving as a time ?ducial,this internal “return”allows one to make a di?erential measurement of target range by referencing the target return time to that of the corner-cube return.If measured in the same way as the target return,the calibration return eliminates systematic errors that are common to both paths.
Because it is necessary to measure both signals in the same manner,the calibration return must be attenuated to the level of the target return—down to the few-photon level.Here,the multiplexed approach o?ers a bene?t over single-photon mode in that one may tune the calibration signal to generate several detected photons per pulse,and in doing so greatly reduce the number of pulses that have no ?ducial reference.For example,if the system is operating at the 0.1photon-per-pulse level,and the calibration is tuned to the same,only about 1%of the target returns will by chance have a calibration return associated with the same pulse.On the other hand,if the system is operating at N photons-per-pulse,the likelihood of detecting no calibration photons in a given pulse is e ?N ,or less than 5%for N =3.Thus one achieves greater assurance that each return shot has a ?ducial time against which to compare.
Complete reliance on the calibration pulse to compute the round-trip travel time ?t =t target ?t cal would tend to produce uncertainties that are √2times worse than the uncertainty of the target or calibration measurements separately.One can improve on this by also measuring the laser ?re time using a fast photodiode and a constant-fraction discriminator to obtain a high signal-to-noise ratio estimate of the laser pulse time [21].Comparison of the photodiode start with the target return does not qualify as a di?erential measurement,since the two are measured in wholly di?erent ways.But one may nonetheless reference the calibration-return-time to the laser-?re-time,using the laser ?re as a precision “anchor ”.The interval between the time anchor and the calibration return may vary as environmental conditions change—such as optical paths,cable delays,and electronics responses.But these changes are likely to be slow,with timescales of several minutes or more.
Under the assumption that the anchor drifts slowly,over short timescales one may reference the time of each calibration photon to this temporary ?ducial.In this way,even calibration returns of only one photon may be properly placed with respect to the departure time of the center of the pulse.Figure 2demonstrates this concept.At top is the system response—a convolution of laser pulse shape,detector jitter,electronics jitter,etc.This represents the distribution one would see emerge after a large number of photon events were recorded.The yellow vertical line indicates the centroid of the distribution.The rising edge of the laser pulse is detected by the fast photodiode,represented here as a thick green line.The placement is misleading since the calibration return arrives long after the photodiode trigger.But in a gedanken manner,one can simply delay the photodiode report until its relative position is as shown here.
Note that the single-shot mean,shown as vertical pink bars in Figure 2,“jumps”around by ～σ/√N ,where σis the standard deviation of the convolved instrument response,and N is the number of photons in a given shot.The photodiode anchor eliminates this noise source,placing the photon events in their proper places (e.g.,shots 3and 11).Over ?ve minutes,operating at 20Hz and 2.5photons per pulse,15,000calibration photon events are recorded.Thus one may build up an accumulated system response distribution with many events per time bin,approaching the black curve at the top of the ?gure.This distribution will di?er from the target return by the small contribution from the fast photodiode jitter,and the generally larger contribution
Figure2:The multi-photon calibration scheme.Fourteen representative shots are displayed at an average rate of4.5photons per shot.Arrows represent individual photons.The vertical pink lines are the means within a shot.The thick green line is the high-precision photodiode start time.The pro?le at top and the vertical yellow line represent the accumulated properties of the photon time distribution.
from the target signature.Because the former is small,the latter may be directly determined by deconvolving the target response pro?le by the calibration response pro?le,provided that multi-photon biases have been properly treated.
In practice,once the o?set between the photodiode anchor and the short timescale calibration ensemble mean(labeled as the pseudo-static o?set in Figure2)is determined,the time of arrival of target return photons may be directly referenced to the high-precision photodiode anchor time for that particular shot.The o?set can be established based on a sliding averaging time—e.g.,a time window of±2minutes about the current time.The character of the o?set may be studied to determine appropriate timescales over which the assumption of slow changes holds,and as a tool for evaluating sources of systematic error.
5Multiplexed Timing Scheme
A key ingredient of the APOLLO timing scheme is a16-channel time-to-digital converter(TDC) made by Phillips Scienti?c(model7186H[22]).At its highest resolution,the TDC measures intervals between0–100ns,at12-bit(25ps)resolution.Our tests show that its RMS jitter is around13ps per event.The TDC is responsible for measuring the time between photon arrival and a?ducial time mark derived from a precision oscillator.Figure3shows the general idea.At some time,in reference to the clock pulse train,the APD is turned(gated)on,after which photons may trigger detection events,and send START pulses to individual TDC channels. Figure3schematically displays six APD channels operating independently.Among the events depicted are four events clustered in a return packet,one sporadic background event,and one
Figure3:Schematic timing scheme.Photon events may occur within the APD gate,which is referenced to the50MHz clock.The time between the photon events and a selected clock pulse after the gate is measured for each channel by the TDC(pink arrow).
channel for which no photons were detected while the gate was on.After the gate is turned o?, the next rising edge of the clock serves as a common STOP pulse to the TDC.Thus each TDC channel measures the interval between the photon event and this?ducial stop time.
Now if we simply know which clock pulse acted as the STOP,and the associated time,we then have all we need to measure the time between pulse departure and pulse return:?t= t STOP2??t TDC2?t STOP1+?t TDC1+t o?set.The clock pulses may be counted continuously,and the counters latched on STOP events for readout.Reseting the counter once per second then ascribes more meaning to the count value:the time within the second.If the one pulse-per-second signal is derived from a clock synchronized to UTC,then the time of photon arrival may be determined to an accuracy of～50ns(GPS?UTC accuracy),and a relative precision of ～35ps(dominated by the APD).
This is essentially the hierarchical scheme employed in APOLLO,composed of three pieces: clock time in seconds,pulse number in20ns intervals,and TDC measurement to25ps resolution. We additionally must predict the appropriate time to turn on the APD so that we catch the lunar return.A block diagram of the digital timing scheme appears in Figure4.Each of the large boxes is clocked by a50MHz digital clock signal.The gate width control and the target delay counter are each counter/comparator schemes that send out a20ns pulse(CLOSE and target START,respectively)when some prescribed count is reached.For example,the counter in the gate width control is reset at either photodiode or target START signals,and counts in20ns intervals(corresponding to the50MHz clock)until the prescribed value,say5,is reached.At that time,the APD gate is commanded to CLOSE,having been OPENed by the START pulse. The target delay counter is never reset,and has a wrap period of5.4seconds.At a gate CLOSE event,the counter value is latched and its data made available to the controlling computer.In the case of a calibration gate event—the one associated with laser?re—the target delay for that pulse is computed and fed via a?rst-in-?rst-out(FIFO)queue back into the target delay counter. When the free-running50MHz counter reaches this count,a target START event is generated.
Also shown are a few time-keeping blocks:the counts-per-second(CPS)block,and the time-within-second(TWS)block.The TWS counter is clocked at50MHz,and reset on the rising edge of the one pulse-per-second(1PPS)signal from the GPS clock(TrueTime XL-DC).Because the 1PPS signal is derived from the50MHz GPS signal,these two are co-phased.By latching the
width set photodiode START
delay value in to APD
data out delay data out STOP Select START Select
1 Pulse per second data out
Figure4:Block diagram of the APOLLO timing scheme.All blocks are clocked at50MHz. value of the TWS counter at a CLOSE event,the exact clock pulse associated with the CLOSE event is unambiguously identi?ed.Thus the time of the gate event within the GPS-clock-de?ned second is identi?ed.The free-running counter,also latched by the CLOSE event,serves as a redundant check of the clock pulse number associated with the gate event.The CPS counter is another redundancy to verify proper counting of all clock pulses.Both latched and reset on the 1PPS signal,this counter should always read50,000,000(minus two,as implemented).
The CLOSE and START signals perform another critical role in addition to the APD gate control.These20ns logic pulses are used to enable comparators whose input is the raw50MHz sine wave from the GPS clock.By bracketing a single positive voltage-swing of the sine wave with the20ns enable pulse,we can send a single clock pulse out of the comparator and to the TDC.In this way,the CLOSE signal slices a chosen clock pulse to act as a STOP,and optionally,the OPEN signal selects a START for the TDC.This latter capability allows us to send a START/STOP pair to the TDC whose time separation is a precise multiple20.0ns—depending on selected gate width—uncertain at the10ps level.The START pulse is simultaneously applied to all16input channels of the TDC,by way of the TEST input.In this way,we can calibrate each channel of the TDC against an absolute standard.At a data rate of1kHz,one may perform a sweep of the TDC at20,40,60,80,and100ns intervals in less than ten seconds,accumulating more than a thousand data points per interval.Rapid and frequent calibration will aid in the understanding and elimination of systematic errors.
The logic block of Figure4also shows various data paths.The inputs are used to set the width of the gate and the count associated with the next target event.The outputs pass the latched counter values.A computer controls this information exchange,recording data and calculating new target delay values.
This general timing scheme relies on multiplexing capability only in the TDC unit.The number of channels employed is irrelevant(until data input/output operations become limiting). APOLLO can easily upgrade to larger detector formats simply by employing additional TDC units.
Figure5:Approximate logic function of the main portion of the TIMER chip.Counters clocked at50MHz trigger events when their counts reach the values set in the adjacent registers.At top left is the gate width control,at top right is the APD gate control,at left center is the target delay generator,and at lower left is the time-within-second recorder.
6Implementation:The APOLLO Command Module
The digital logic associated with the APOLLO timing scheme has been implemented on a custom CAMAC module called the APOLLO Command Module(ACM).The ACM is little more than two Altera programmable chips(MAX-7000AE series)interfacing to the CAMAC dataway and to various external hardware.The TIMER chip performs the timing logic,and contains the elements schematically depicted in Figure4.The CAMAC chip controls the CAMAC interface, parsing commands,passing data,and setting the operating state of the TIMER chip.The chips are re-programmable in situ up to1000times.Thus the details of the logic functions may be changed via software,without needing access to the hardware.
A schematic example of the core elements in the TIMER chip(those represented in block form in Figure4)is shown in Figure5,though the details of the real implementation di?er.The function of the ACM extends beyond that of counting clock pulses and controlling the APD gate. In total,the ACM performs the following tasks:
?keeping track of time(counting clock pulses and registering signi?cant epochs),
?opening the APD gate in anticipation of photon arrival,
?controlling the duration of the APD gate,
??ring the laser in synchronization with the rotating transmit/receive switch,
?telling the50MHz stage when to send a STOP pulse to the TDC,
Figure6:Oscilloscope trace of the ACM output,showing the OPEN signal and the GATE signal, about6ns later.The crossing at top is confusing,but the OPEN pulse is the shorter(20ns) pulse that rises?rst.
?sending a time request to the GPS clock once per second,
?controlling entry into“stare”mode,allowing the APD to sense stars,
?blocking the laser when safety demands,
?and calibrating the TDC with pulses separated by precise multiples of20.00ns.
In addition,the ACM has safety features built in to prevent accidental damage to the APD or laser.There are also?ve as-yet unassigned ports,con?gurable as input or output,that may be used for future purposes.Potential uses include deploying corner cubes,setting the state of variable attenuators,etc.
The CAMAC communication rate is relatively high-speed,with a typical command duration of 1.5–2.0μs.Test trials of the ACM controlled from a Linux machine running the stock RedHat7.1 kernel(2.4.7)showed that the computer was able to keep up with the data rate(about12CAMAC commands per cycle)with no dropouts at rates of300Hz.Data dropouts were less than0.5% up to about20kHz,becoming50%at～40kHz.It is at this point that the duration of the CAMAC command sequence—mostly reading data from the TIMER chip—exceeds the cycle time.These tests are somewhat unrealistic for the?nal control loop,because the roughly20 commands necessary to read and reset all TDC channels were not included.Nonetheless,this scheme is capable of operation at several kHz in a practical ranging application.The few dropouts that do occur can be eliminated or reduced by applying low-latency patches to the Linux kernel, or by running Real-Time Linux,or some other low-latency operating system.
Figure6shows the appearance of ACM output pulses.In particular,the APD gate and the OPEN request pulses are shown.Each is a multiple of20ns in duration.The OPEN request is generated by the target delay counter,shortly followed by the gate pulse.
A picture of the“business end”of the ACM is shown in Figure7.The CAMAC card-edge connector and input/output bu?ers are left out of this picture at bottom.At upper left is the
Figure7:The APOLLO Command Module,showing the Altera programmable chips,and the many input/output connectors.
TIMER chip,with10,000gates,512macrocells,and7ns propagation delay.At bottom is the CAMAC chip,with5,000gates,256macrocells,and10ns propagation delay.Along the top, one can see the row of22LEMO-style connectors for interface to the rest of the system.The rightmost?ve connectors are jumper-con?gurable to act as outputs from or inputs to the TIMER chip.
The central role that the ACM plays in the APOLLO scheme may be appreciated by exami-nation of Figure8,showing the total APOLLO electronics system.The number of lines?owing into and out of the ACM—also evident from the number of connectors in Figure7—would lead any physiologist to conclude that this must be the heart of the system.Another key interface is provided by the“booster stage,”which multiplies the10MHz signal from the GPS clock by ?ve,processes the selection of START and STOP pulses,and converts the NIM signal from the photodiode discriminator to an ECL(emitter-coupled logic)signal for the TDC.
7Project Status
The color-coding of Figure8serves to show the status of integration of the APOLLO system. Most of the core system is in place.The ACM,having recently been completed,is undergoing initial testing.The APD circuitry is still in active development,and will likely continue to evolve up to the point of installation.Some of the peripheral equipment that is not yet needed has not yet been explored/selected.
The APDs have been tested,and we have concluded that their intrinsic jitter is in the neigh-borhood of30ps(see[19]).The APOLLO error budget therefore looks very promising,as summarized in Table1.The single-photon RMS of the APOLLO system is around8mm,so that1mm precision may be statistically achieved in<100photons.Of course,the lunar re-?ector arrays introduce a large statistical spread depending on orientation,such that our photon requirement becomes roughly1000per normal point.If we detect one photon per pulse at20Hz, we will achieve this goal in less than one minute.In practice,?ve minutes per re?ector is likely
Table1:APOLLO Error Budget
Statistical Error Source RMS Error(ps)One-way Error(mm)
Laser Pulse(95ps FWHM)406
APD Jitter30 4.5
TDC Jitter15 2.2
50MHz Freq.Reference71
APOLLO System Total528
Lunar Retrore?ector Array80–23012–35
Total Error per Photon100–24014–37
to be su?cient.
The laser[23]has been delivered to Apache Point,and awaits installation on the telescope. Because the laser will be mounted on the telescope in the cold ambient environment,we are currently spending our time developing the appropriate thermal protections—both to the laser and to the dome environment.
Much remains to do before we acquire our?rst lunar data.We are optimistic that2003will bring many happy returns.See the APOLLO website for updates on our status[24]. Acknowledgements
We thank Tim van Wechel for his substantial design work on the APOLLO Command Module, and Allan Myers for its excellent fabrication.We also thank the Center for Experimental Nuclear Physics and Astrophysics(CENPA)for their donation of resources.Jesse Angle and Dan Miller also contributed to the project during the last year.APOLLO is supported by a grant from NASA.
e d , w o r
k
s
p m e n
t
s
y
e
t
F
i g u r e 8:A P O L L O b l o c k d i a g r a m s h o w i n g t h e p r i m a r y e l e c t r o n i c s c o m p o n e n t s .C o l o r -c o d i n g i n d i c a t e s c u r r e n t s t a t u s .
References
[1]Bender,P.L.,Currie,D.G.,Dicke,R.H.,Eckhardt,D.H.,Fal ler,J.E.,Kaula,W.M.,
Mullholland,J.D.,Plotkin,H.H.,Poultney,S.K.,Silverberg,E.C.,Wilkinson,D.T., Williams,J.G.,&Alley,C.O.,“The Lunar Laser Ranging Experiment,”Science,182,229, (1973)
[2]Dickey,J.O.,Bender,P.L.,Faller,J.E.,Newhall,X.X.,Rick lefs,R.L.,Ries,J.G.,Shelus,
P.J.,Veillet,C.,Whipple,A.L.,Wiant,J.R.,Williams,J.G.,&Yoder,C.F.,“Lunar Laser Ranging:A Continuing Legacy of the Apollo Program,”Science,265,482,(1994)
[3]Williams,J.G.,Boggs,D.H.,Dickey,J.O.,&Folkner,W.M.,“Lunar Tests of the Gravita-
tional Physics,”Proceedings of the Ninth Marcel Grossman Meeting,Rome,Italy,June2000, World Scienti?c Publications,R.Jantzen,ed.,(2001)
[4]Anderson,J.D.,&Williams,J.G.,“Long-Range Tests of the Equivalence Principle,”Class.
Quantum Grav.,18,2447,(2001)
[5]https://www.doczj.com/doc/242563717.html,/rlr/mlrs.html
[6]Murphy,T.W.,Strasburg,J.D.,Stubbs,C.W.,Adelberger,E.G.,Angle,J.,Nordtvedt,
K.,Williams,J.G.,Dickey,J.O.,&Gillespie,B.,“The Apache Point Observatory Lunar Laser-ranging Operation(APOLLO),”12th International Laser Ranging Workshop,Matera, Italy,(2000).
[7]Nordtvedt,K.,“Equivalence Principle for Massive Bodies.I.Phenomenology,”Physical Re-
view,169,1014,(1968)
[8]Nordtvedt,K.,“Equivalence Principle for Massive Bodies.II.Theory,”Physical Review,169,
1017,(1968)
[9]Damour,T.,“Testing the equivalence principle:why and how?,”Class.Quantum Grav.,13,
A33,(1996)
[10]Bae?ler,S.,Heckel,B.R.,Adelberger,E.G.,Gundlach,J.H.,Schmidt,U.,&Swanson,
H.E.,“Improved Test of the Equivalence Principle for Gravitational Self-Energy,”Physical
Review Letters,83,3585,(1999)
[11]Damour,T.,&Nordtvedt,K.,“Tensor-scalar cosmological models and their relaxation
toward general relativity,”Phys.Rev.D,48,3436,(1993)
[12]Dvali,G.,Gruzinov,A.,&Zaldarriaga,M.,“The Accelerated Universe and the Moon,”
LANL preprint archive,https://www.doczj.com/doc/242563717.html,/abs/hep-ph/0212069,5Dec.,(2002)
[13]Lue,A.,&Starkman,G.,“Gravitational Leakage into Extra Dimensions:Probing Dark En-
ergy Using Local Gravity,”LANL preprint archive,https://www.doczj.com/doc/242563717.html,/abs/astro-ph/0212083, 11Dec.,(2002)
[14]Williams,J.G.,Newhall,X.X.,&Dickey,J.O.,“Relativity parameters determined from
lunar laser ranging,”Physical Review D,53,6730,(1996)
[15]Samain,E.,Mangin,J.F.,Veillet,C.,Torre,J.M.,Fridelence,P.,et al.,“Millimetric Lunar
Laser Ranging at OCA(Observatoire de la C?o te d’Azur,”Astronomy&Astrophysics Suppl.
Ser.,130,235,(1998)
[16]Silverberg,E.C.,private communiucation(2002)
[17]Wiant,J.R.,private communiucation(2002)
[18]Samain,E.,private communiucation(2002)
[19]Strasburg,J.D.,et al.,“The advantages of Avalanche Photodiode(APD)arrays in laser
ranging applications,”13th International Laser Ranging Workshop,Washington,D.C.,USA, (2002).
[20]Nordtvedt,K.,“Optimizing the observation schedule for tests of gravity in lunar laser ranging
and similar experiments,”Classical Quantum Gravity,15,3363,(1998)
[21]Using Hamamatsu G4176photodiode,Picosecond Pulse Labs Model5545bias tee,and Ortec
9327ampli?er/constant-fraction-discriminator.
[22]https://www.doczj.com/doc/242563717.html,/phisci1.htm
[23]https://www.doczj.com/doc/242563717.html,/leopard.html
[24]https://www.doczj.com/doc/242563717.html,/tmurphy/apollo/apollo.html
LinuxApacheWeb服务器配置教程 Linux阿帕奇网络服务器配置教程 阿帕奇的主要特点 根据著名的万维网服务器研究公司进行的一项调查，全世界50%以上的万维网服务器使用阿帕奇，排名世界第一。 阿帕奇的出生非常戏剧化。当NCSA万维网服务器项目停止时，那些使用NCSA万维网服务器的人开始用他们的补丁来交换服务器，他们很快意识到有必要建立一个论坛来管理这些补丁。就这样，阿帕奇集团诞生了，后来这个集团在NCSA的基础上建立了阿帕奇。 阿帕奇的主要特点是: 。可以在所有计算机平台上运行； 。支持最新的HTTP 1.1协议； 简单而强大的基于文件的配置； 。支持通用网关接口CGI 。支持虚拟主机； 。支持HTTP认证； 。集成的Perl脚本编程语言； 。集成代理服务器； 。拥有可定制的服务器日志；。支持服务器端包含命令。支持安全套接字层。用户会话过程的跟踪能力；支持FastCGI。支持Java小服务程序。 安装Apache流程
安装Apache 接下来，我们将开始征服阿帕奇的漫长旅程。我们将一步一步地学习使用Apache，从介绍到掌握，通过需求的一步一步的例子。 系统需求 运行Apache不需要太多的计算资源。它运行良好的Linux系统有6-10MB的硬盘空间和8 MB的内存。然而，单独运行Apache可能不是您想要做的。更有可能的是，您希望运行Apache来提供WWW服务，启动CGI流程，并充分利用WWW所能提供的所有惊人功能。在这种情况下，您需要提供额外的磁盘空间和内存空间来反映负载要求。换句话说，它不需要太多的系统资源来启动WWW服务，但是它需要更多的系统资源来为大量的客户提供服务。获取软件 你可以呆在 错误日志命令用于指定错误日志文件名和路径。 命令格式:错误日志[日志文件名] 示例:错误日志/var/ srm.conf Srm.conf是一个资源配置文件，它告诉服务器您想在WWW站点上提供什么资源，在哪里以及如何提供这些资源。 DocumentRoot命令指定主文档的地址。 命令格式:文档根[路径] 示例:文档根目录/主页/ UserDir命令，用于指定个人主页的位置。如果你有一个用户测试，
在Linux下配置Apache服务器 一、实验目的 完成本次实训，将能够: ●配置基本的Apache服务器 ●配置个人用户Web站点。 ●配置虚拟目录别名功能。 ●配置主机访问控制。 ●配置用户身份验证功能.。 ●配置基于IP地址的虚拟主机. 二、实验环境 1、RedHat Linux4AS. 2、Apache 2.0 三、实验内容 1.配置基本的Apache服务器 2.配置个人用户Web站点。 3.配置虚拟目录别名功能。 4.配置主机访问控制。 5.配置用户身份验证功能.。 6.配置基于IP地址的虚拟主机。 四、实验要求 在Linux操作系统下配置Apache服务器。 五、注意事项 1.在修配置文件下注意区分大小写、空格。 2.在每次重新开机后都必须启动Apachec服务器。 3.在每次修改完主配置文件后保存起来，必须重启Apachec服务器，如果不重启会 导致配置无效，最终导致实验失败。 六、实验步骤 1、检测是否安装了Apache软件包: A、首先为服务器网卡添加一个固定的IP地址。 B、在Web浏览器的地址栏中输入本机的IP地址，若出现Test Page测试页面（该 网页文件的默认路径为var/www/html/index.html）如下图1所示就说明Apache 已安装并已启动。
另一种方法是使用如下命令查看系统是否已经安装了Apache软件包： [root@rhe14~]# rpm –aq | grep httpd Httpd-suexec-2.0.52-9.ent Httpd-manual-2.0.52-9.ent System-config-httpd-1.3.1-1 Httpd-devel-2.0.52-9.ent 出现以上内容表明了系统已安装Apache软件包。 2、安装Apache软件包 超级用户（root）在图形界面下选择“应用程序”|“系统设置”|“添加/删除应用程序”命令，选择“万维网服务器”软件包组，在单击“更新”按钮就可以安装与Apache相关的软件包。 3、Apache的基本配置 （1）打开终端输入[root@rhe14~]# /etc/rc.d/init.d/httpd start //启动Apache 或者 [root@rhe14~]# apachectl start //启动Apache [root@rhe14~]# apachectl stop //停止Apache服务 [root@rhe14~]# apachectl restart //重启Apache服务 [root@rhe14~]# apachectl configtest //测试Apache服务器配置语法（2）在httpd.conf将Apache的基本配置参数修改、将一些注释的语句取消注释，或将某些不需要的参数注释掉。 （3）将包括index.html在内的相关网页文件复制到指定的Web站点根目下（var/www/html/index.html） （4）重启httpd进程 (5) 在Web浏览器下输入配置的ip地址出现如下图2，那表明基本配置成功了:
重要提示 Web服务器包括apache的安装部署和W AS7 Plugin安装部署两部分，如果的websphere应用服务器使用非集群模式，plugin则不需要安装，只需配置本文2.1章节内容，如果websphere 应用服务器使用群集模式，则需要按照本文2.2章节进行plugin安装配置。 1 Apache安装 Apache的安装和配置现在可以采用脚本自动化安装，脚本就是139ftp上的 apache_install_script.sh 请下载到web服务器中，并执行即可。 注意：在执行脚本安装前请确认web服务器的/opt/apache下没有安装过apache，并且web 服务器能上外网（能ping通https://www.doczj.com/doc/242563717.html,） 成功安装apache并测试通过后即可直接继续本文第二章节Was7 Plugin安装 在root下进行root进入方法#su 然后输入密码 1.1 准备安装 关闭系统自带的web服务： #chkconfig httpd off 在线安装gcc #yum install gcc cc 下载并解压安装程序： #cd ~/ #wget https://www.doczj.com/doc/242563717.html,/httpd/httpd-2.2.15.tar.gz #tar –zxvf httpd-2.2.15.tar.gz -C /usr/src Web 服务器安装部署手册 Page 4 of 21 1.2 安装Apache Web Server 进入源码目录： #cd /usr/src/httpd-2.2.15 编译源文件： #./configure //(安装到默认目录) 形成安装文件: #make 安装程序： #make install 1.3 验证安装 进入安装后目录： # cd /opt/apache/apache-2.2.15/bin 检查进程模式： #./apachectl –l Compiled in modules: core.c worker.c http_core.c mod_so.c 启动Apache Web Server：
【实验8】Apache服务器的安装和配置 一、实验目的： 1.掌握Apache Web服务器的安装和配置。 2.使用虚拟主机在同一台服务器上架设多个网站。 二、【实验环境】 1.虚拟机软件VM Ware 6.0，Redhat Enterprise Linux虚拟机或光盘镜像文 件。 2.2台以上机器组成的局域网。 三、【实验原理】 （一）Apache服务简介 Apache是世界使用排名第一的Web服务器软件。它可以运行在几乎所有广泛使用的计算机平台上。 Apache源于NCSAhttpd服务器，经过多次修改，成为世界上最流行的Web 服务器软件之一。Apache取自“a patchy server”的读音，意思是充满补丁的服务器，因为它是自由软件，所以不断有人来为它开发新的功能、新的特性、修改原来的缺陷。Apache的特点是简单、速度快、性能稳定，并可做代理服务器来使用。 （二）虚拟主机 所谓虚拟主机，也叫“网站空间”就是把一台运行在互联网上的服务器划分成 多个“虚拟”的服务器，每一个虚拟主机都具有独立的域名和完整的Internet服务 器（支持WWW、FTP、E-mail等）功能。一台服务器上的不同虚拟主机是各自 独立的，并由用户自行管理。 虚拟主机技术是互联网服务器采用的节省服务器硬体成本的技术，虚拟主机 技术主要应用于HTTP服务，将一台服务器的某项或者全部服务内容逻辑划分 为多个服务单位，对外表现为多个服务器，从而充分利用服务器硬体资源。如果 划分是系统级别的，则称为虚拟服务器。
（三）Linux中虚拟主机的分类 1、基于IP地址的虚拟主机 如果某公司有多个独立的IP地址可用，那么可以用不同的IP地址来配置虚拟主机。 2、基于端口的虚拟主机 如果只有一个IP地址，但是要架设多个站点，可以使用端口来区分，每个端口对应一个站点。这样配置的话，用户在访问的时候必须在 URL中指明端口号才能访问相应的网站。 3、基于名称的虚拟主机 使用基于IP地址或者端口的虚拟主机，能够配置的站点数目有限，而使用基于名称的虚拟主机，可以配置任意数目的虚拟主机，而不需要 额外的IP地址，也不需要修改端口号。 四、实验步骤 本实验请勿使用【系统】→【管理】→【服务器设置】中的【HTTPD】工具来配置，否则后果自负！ （一）Apache服务器的启动 1、测试是否已安装Apache服务器： [root@localhost ~]#rpm –qa httpd 2、启动Apache服务器： [root@localhost ~]#service httpd start （二）基于端口的虚拟主机的配置 1、在/etc/httpd目录中，建立一个名为vhostconf.d的子目录，用来存放虚拟 主机的配置文件。 2、在/var/www目录中，建立一个名为websites的子目录，用于存放网站源 文件；在website目录下再建立ipvhost1和ipvhost2文件夹，用于区分各 个站点。
Linux下搭建web服务器 Apache源于NCSAhttpd服务器，经过多次修改，成为世界上最流行的 Web服务器软件之一。Apache取自“a patchy server”的读音，意 思是充满补丁的服务器，因为它是自由软件，所以不断有人来为它开发新的功能、新的特性、修改原来的缺陷。Apache的特点是简单、 速度快、性能稳定，并可做代理服务器来使用。 本来它只用于小型或试验 Internet网络，后来逐步扩充到各种Unix 系统中，尤其对Linux的支持相当完美。Apache有多种产品，可以支持SSL技术，支持多个虚拟主机。Apache是以进程为基础的结构，进程要比线程消耗更多的系统开支，不太适合于多处理器环境，因此，在一个Apache Web站点扩容时，通常是增加服务器或扩充群集节点 而不是增加处理器。到目前为止Apache仍然是世界上用的最多的Web 服务器，市场占有率达60%左右。世界上很多著名的网站如Amazon.c om、Yahoo!、W3 Consortium、Financial Times等都是Apache的产物，它的成功之处主要在于它的源代码开放、有一支开放的开发队伍、支持跨平台的应用（可以运行在几乎所有的Unix、Windows、Linux 系统平台上）以及它的可移植性等方面。 Apache的诞生极富有戏剧性。当NCSA WWW服务器项目停顿后，那些 使用NCSA WWW服务器的人们开始交换他们用于该服务器的补丁程序，他们也很快认识到成立管理这些补丁程序的论坛是必要的。就这样，诞生了Apache Group，后来这个团体在NCSA的基础上创建了Apache。 Apache的主要特征是： 可以运行上所有计算机平台； 支持最新的H TT P1.1协议； 简单而强有力的基于文件的配置； 支持通用网关接口CGI； 支持虚拟主机； 支持H TT P认证； 集成P erl脚本编程语言；
Apache服务器的安装与配置 一、安装Apache 双击可执行文件apache_1.3.33-win32-x86-no_src.exe，将Apache服务器软件安装至C:\Apache目录下。 二、设置C:\apache\conf\httpd.donf文件 修改Apache的核心配置文件c:\apache\conf\httpd.conf（说明一点：“#”为Apache的注释符号）。修改方法如下： 1、寻找到ServerName。这里定义你的域名。这样，当Apache Server运行时，你可以在浏览器中访问自己的站点。如果前面有#，记得删除它。 2、寻找到ServerAdmin。这里输入你的E-Mail地址。 （以上两条在安装时应该已经配置好了，所以不必改动，这里介绍一下，主要是为了日后的修改） 3、寻找到。向下有一句Options，去掉后面所有的参数，加一个All（注意区分大小写！A 大写，两个l小写。下同。）；接着还有一句Allow Override，也同样去掉后面所有的参数，加一个All。
1、如何设置请求等待时间 在httpd.conf里面设置： TimeOut n 其中n为整数，单位是秒。 设置这个TimeOut适用于三种情况： 2、如何接收一个get请求的总时间 接收一个post和put请求的TCP包之间的时间 TCP包传输中的响应（ack）时间间隔 3、如何使得apache监听在特定的端口 修改httpd.conf里面关于Listen的选项，例如： Listen 8000 是使apache监听在8000端口 而如果要同时指定监听端口和监听地址，可以使用： Listen 192.170.2.1:80 Listen 192.170.2.5:8000 这样就使得apache同时监听在192.170.2.1的80端口和192.170.2.5的8000端口。 当然也可以在httpd.conf里面设置： Port 80 这样来实现类似的效果。 4、如何设置apache的最大空闲进程数 修改httpd.conf，在里面设置： MaxSpareServers n 其中n是一个整数。这样当空闲进程超过n的时候，apache主进程会杀掉多余的空闲进程而保持空闲进程在n，节省了系统资源。如果在一个apache非常繁忙的站点调节这个参数才是必要的，但是在任何时候把这个参数调到很大都不是一个好主意。 同时也可以设置： MinSpareServers n 来限制最少空闲进程数目来加快反应速度。 5、apache如何设置启动时的子服务进程个数 在httpd.conf里面设置： StartServers 5 这样启动apache后就有5个空闲子进程等待接受请求。 也可以参考MinSpareServers和MaxSpareServers设置。 6、如何在apache中设置每个连接的最大请求数 在httpd.conf里面设置： MaxKeepAliveRequests 100 这样就能保证在一个连接中，如果同时请求数达到100就不再响应这个连接的新请求，保证了系统资源不会被某个连接大量占用。但是在实际配置中要求尽量把这个数值调高来获得较高的系统性能。 7、如何在apache中设置session的持续时间 在apache1.2以上的版本中，可以在httpd.conf里面设置： KeepAlive on KeepAliveTimeout 15 这样就能限制每个session的保持时间是15秒。session的使用可以使得很多请求都可以通过同一个tcp
Apache 安装后启动然后配置即可.记得开启80 端口. iptables –I INPUT –p tcp –dport 80 –j ACCEPT 开启tcp 80 端口 apache 的配置文件: httpd.conf 路径: /etc/httpd/conf/httpd.conf DocumentRoot “/var/www/html”设置主目录的路径 DirectoryIndex index.html index.html.var 设置默认主文档,中间用空格格开 Listen 80 Listen 192.168.1.1:80 设置apache监听的IP地址和端口号,可添加多个 ServerRoot “/etc/httpd”设置相对根目录的路径(存放配置文件和日志文件) ErrorLog Logs/error_log 设置错误日志存放路径 CustomLog Logs/access_log combined (日志格式) 设置访问日志存放路径 如果日志文件存放路径不是以”/”开头,则意味着该路径相对于ServerRoot 的相对路径. ServerAdmin 邮箱地址{设置管理员的E-mail地址 ServerName FQDN名或IP地址{设置服务器主机名 由于Apache默认字符集为西欧(UTF-8),所以客户端访问中文网页时会出现乱码. 将语句“AddDefaultCharset UTF-8”改为“AddDefaultCharset GB2312”方可解决,不过要重新启动Apache服务. 修改完默认字符集后,客户端如需访问,要先清空浏览器的缓存.
创建虚拟目录,添加Alias语句即可 Alias /ftp “/var/ftp”Alias 虚拟目录名物理路径
A p a c h e服务器配置毕 业设计 目录 摘要 ................................................................................................................ 错误！未定义书签。 1 综述 (1) 1.1 架设WWW网站的意义 (1) 1.2 WWW的工作原理 (1) 1.3 在Linux下构建WWW服务器 (2) 1.3.1 关于硬件配置 (2) 1.3.2 将linux用作www服务器 (2) 2 Red Hat Linux的安装与使用 (4) 2.1 Red Hat Linux9.0简介 (4) 2.1.1 Red Hat Linux的网络功能 (6) 2.1.2 Red Hat Linux的文件类型 (9) 2.2 Red Hat Linux9.0的安装和配置 (11) 2.2.1 合理划分分区 (11) 2.2.2 了解相关信息 (12) 2.2.3 图形化安装过程 (13) 3 WWW服务器的建立 (27) 3.1 Apache的体系结构及性能 (27) 3.1.1 Apache的体系结构 (27) 3.1.2 Apache性能简介 (28) 3.2 配置并启动Apache (31) 3.2.1 配置文件httpd.conf (31) 3.2.2 Apache服务的安装、启动与停止 (32) 3.3 设置用户个人主页 (33) 3.3.1设置Linux系统用户个人主页的目录 (33) 3.3.2设置用户个人主页所在目录的访问权限 (33) 3.4 设置虚拟主机 (34) 3.4.1 配置DNS (34) 4 建立和完善WWW站点 (42) 4.1建立安全传输的WWW站点 (42) 4.1.1认识SSL安全协议 (42) 4.1.2 维护站点安全性应注意的问题 (42) 5 结论 (44) 致谢 ................................................................................................................ 错误！未定义书签。参考文献.. (45) 8
Apache服务器配置安全规范以及其缺陷！正如我们前言所说尽管Apache服务器应用最为广泛，设计上非常安全的程序。但是同其它应用程序一样，Apache也存在安全缺陷。毕竟它是完全源代码，Apache服务器的安全缺陷主要是使用HTTP 协议进行的拒绝服务攻击(denial of service)、缓冲区溢出攻击以及被攻击者获得root权限三缺陷和最新的恶意的攻击者进行拒绝服务(DoS)攻击。合理的网络配置能够保护Apache服务器免遭多种攻击。我们来介绍一下主要的安全缺陷。主要安全缺陷（1）使用HTTP协议进行的拒绝服务攻击(denial of service)的安全缺陷这种方法攻击者会通过某些手段使服务器拒绝对HTTP应答。这样会使Apache对系统资源(CPU时间和内存)需求的剧增，最终造成Apache系统变慢甚至完全瘫痪。（2）缓冲区溢出的安全缺陷该方法攻击者利用程序编写的一些缺陷，使程序偏离正常的流程。程序使用静态分配的内存保存请求数据，攻击者就可以发送一个超长请求使缓冲区溢出。（3）被攻击者获得root权限的安全缺陷该安全缺陷主要是因为Apache服务器一般以root权限运行(父进程)，攻击者会通过它获得root权限，进而控制整个Apache系统。（4）恶意的攻击者进行拒绝服务(DoS)攻击的安全缺陷这个最新在6月17日发现的漏洞，它主要是存在于Apache的chunk encoding中，这是一个HTTP协议定义的用于接受web用户所提交数据的功能。所有说使用最高和最新安全版本对于加强Apache Web服务器的安全是至关重要的。正确维护和配置Apache服务器虽然Apache服务器的开发者非常注重安全性，由于Apache服务器其庞大的项目，难免会存在安全隐患。正确维护和配置Apache WEB服务器就很重要了。我们应注意的一些问题：（1）Apache服务器配置文件Apache Web服务器主要有三个配置文件，位于 /usr/local/apache/conf目录下。这三个文件是：httpd.conf-----主配置文件srm.conf------填加资源文件access.conf---设置文件的访问权限（2）Apache服务器的目录安全认证在Apache Server中是允许使用 .htaccess做目录安全保护的，欲读取这保护的目录需要先键入正确用户帐号与密码。这样可做为专门管理网页存放的目录或做为会员区等。在保护的目录放置一个档案，档名为.htaccss。AuthName 会员专区 AuthType BasicAuthUserFile /var/tmp/xxx.pw -----把password放在网站外 require valid-user 到apache/bin目录，建password档 % ./htpasswd -c /var/tmp/xxx.pw username1 -----第一次建档要用参数-c % /htpasswd /var/tmp/xxx.pw username2 这样就可以保护目录内的内容，进入要用合法的用户。注：采用了Apache内附的模组。也可以采用在httpd.conf中加入：options indexes followsymlinks allowoverride authconfig order allow,deny allow from all （3）Apache服务器访问控制我们就要看三个配置文件中的第三个文件了，即access.conf文件，它包含一些指令控制允许什么用户访问Apache目录。应该把deny from all设为初始化指令，再使用allow from指令打开访问权限。order deny,allowdeny from allallow from https://www.doczj.com/doc/242563717.html, 设置允许来自某个域、IP地址或者IP段的访问。（4）Apache服务器的密码保护问题我们再使 用.htaccess文件把某个目录的访问权限赋予某个用户。系统管理员需要在httpd.conf或者rm.conf文件中使用 AccessFileName指令打开目录的访问控制。如：AuthName PrivateFilesAuthType BasicAuthUserFile /path/to/httpd/usersrequire Phoenix# htpasswd -c /path/to/httpd/users Phoenix设置Apache服务器的WEB和文件服务器我们在Apache服务器上存放WEB 服务器的文件，供用户访问，并设置/home/ftp/pub目录为文件存放区域，用
实验报告---apache服务器配置 一、实验目的： Apache源于NCSAhttpd服务器，经过多次修改，成为世界上最流行的Web 服务器软件之一。本来它只用于小型或试验Internet网络，后来逐步扩充到各种Unix系统中，尤其对Linux的支持相当完美。Apache有多种产品，可以支持SSL技术，支持多个虚拟主机。Apache的开发遵循GPL协议，由全球志愿者一起开发并维护。它的成功之处主要在于它的源代码开放、有一支开放的开发队伍、支持跨平台的应用（可以运行在几乎所有的Unix、Windows、Linux系统平台上）以及它的可移植性等方面。 支持最新的HTTP1.1通信协议；强大的可配置性和可扩展性；提供全部源代码和不受限制的使用许可；通过第三方模块可以支持Java Servlets；广泛的应用和支持多种平台。 因此做这个实验的目的是为了熟悉apache服务器的配置，通过实验来加深对它的了解。 二、实验内容： （一）、Apache服务器的安装、启动与停止 （二）、配置用户个人主页 （三）、配置虚拟主机：a、创建基于IP地址的虚拟主机 b、创建基于域名的虚拟主机 三、实验步骤： （一）Apache服务器的安装、启动及访问 1.检验apache服务的软件包是否安装，默认情况下是没有安装的，因此需安装。 # rpm -qa|grep httpd //检验软件包是否安装 # mount /dev/cdrom /mnt/cdrom //加载光驱 # cd /mnt/cdrom/Server //进入光盘的Server目录 # rpm -ivh postgresql-libs-8.1.18-2.e15_4.1.i386.rpm //安装所需要的RPM包 # rpm -ivh apr-1.2.7-11.e15_3.1.i386.rpm # rpm -ivh apr-util-1.2.7-11.e15.i386.rpm # rpm -ivh httpd-2.0.40-21.i386.rpm 注意：由于安装apache软件包有依赖关系，因此按照上述顺序进行安装 2.检验网络的连通性
Apache+WebSphere 服务器部署方案 版本号：V1.1 2011年9月
目录 重要提示 (2) 1Linux 安装 (2) 1.1安装全部程序 (2) 1.2安装源设置 (2) 2Apache安装 (3) 2.1准备安装 (3) 2.2安装Apache Web Server (3) 2.3验证安装 (4) 2.4编译代理/反向代理模块 (5) 2.5配置代理/反向代理模块 (6) 2.6创建配置文件crossdomain.xml (8) 2.7Apache中文乱码问题解决 (8) 2.8配置RewriteRule (11) 3Was7 Plugin安装 (11) 3.1Websphere常用命令 (12) 3.2JDNI配置 (12) 3.3应用服务器单机版配置 (16) 3.4应用服务器集群版配置 (16) 4常见问题 (23) 4.1环境问题 (23) 4.2应用问题 (24)
重要提示 Web服务器包括apache的安装部署和WAS7 Plugin安装部署两部分，如果的websphere应用服务器使用非集群模式，plugin则不需要安装，只需配置本文2.1章节内容，如果websphere应用服务器使用群集模式，则需要按照本文2.2章节进行plugin安装配置。 1Linux 安装 1.1 系统安装 1.根分区磁盘容量应大于20G 2.推荐安装所有应用程序 1.2 安装源设置 图中黑色边框指示设置系统安装源，尤其在服务器无法连接外网时需要使用光盘作为安装源
2Apache安装 2.1 准备安装 关闭系统自带的web服务： #chkconfig httpd off 安装gcc(适用于系统没有安装gcc时) #yum install gcc* （yum install gcc-c++ libstdc++-devel） 出现y/n提示时，输入y 下载并解压安装程序： #cd ~/ #wget https://www.doczj.com/doc/242563717.html,/httpd/httpd-2.2.15.tar.gz #tar –zxvf httpd-2.2.15.tar.gz -C /usr/src 把包解压到/usr/src 2.2 安装Apache Web Server 进入源码目录： #cd /usr/src/httpd-2.2.15 配置编译文件： #./configure --prefix=/opt/apache --enable-so --enable-mods-shared=most --with-mpm=worker 编译文件: #make
web服务器的发送和接受 可以通过httpwatch来了解 https://www.doczj.com/doc/242563717.html, cmd中的操作 1.apache其实就输一个软件，apache安装目录下的bin\httpd.exe 先进入到apache\bin目录 httpd -k start[shutdown][restart] 2.如果你希望在任何一个目录下都可以运行我们的httpd.exe则需要做一个环境变量的设置 重新进入控制台cmd 关于端口： apache默认在80端口监听 一台机器可以有：1-65535号端口（2个字节(256*256)） 查看系统正在监听的端口：netstat -an或netstat -nab（发现是哪个程序在监听） 一台机器的一个端口只能被一个程序监听，一个程序可以监听多个端口 端口分为有名端口1-1024，一般别去用 apache修改配置端口 配置文档位置：apache2.2/conf/http.conf 1.添加Listen 8080 2.重新启动apache apache的目录 /bin 该目录用于存放apache常用的命令，如httpd /cgi-bin 该目录用于存放linux/unix下常用的命令 /conf 该目录用于存放配置文件，如httpd.conf /error 该目录用于存放apache启动关闭的错误 /htdocs 该目录用于存放我们的默认站点的文件夹 /icons 该目录用于存放图标 /logs 该目录用于存放apache的相关日志，如error.log,install.log /manual 该目录用于存放apache的手册 /modules 该目录用于存放apache的模块.so的文件 apache是基于模块化设计的，核心代码并不多，大多数的功能都被分散到各个模块中运行机制： MPM:multi processing modules:多重处理模块 APR:Apache Portable Runtime:可移植运行库 生命周期： 启动，配置=》模块初始化=》子进程初始化=》请求循环=》子进程结束 启动阶段===============================》运行阶段============= Apache配置虚拟目录(默认是htdocs) 1.配置虚拟目录在apache的conf目录下的httpd.conf的
在Linux虚拟机下配置apache构建web服务器 2009年07月02日星期四 00:33 实现目的: 在Fedora10.0下架设Apache服务器，为Windows提供web访问服务。实现不同用户（test1,test2，mm1，mm2）的不同访问权限。并且实现基于IP（192.168.1.6与192.168.1.119）和端口（192.168.1.6：80与192.168.1.6：8090）的虚拟主机功能。 实现步骤： 安装好Fedora7.0后，系统已经安装好了Apache服务。路径为 /etc/httpd 其中/etc/httpd/conf/httpd.conf为Apache服务的主配置文件，下面进行配置。ServerRoot "/etc/httpd" //指定Apache服务的启动路径 Listen 192.168.1.6:80 //启动侦听端口 Listen 192.168.1.6:8090 //启动基于端口8090的虚拟主机的侦听 Listen 192.168.1.119:80 //启动基于端口80的虚拟主机的侦听 User apache Group apache //指明启动Apache服务的用户和组 ServerAdmin ccx193@https://www.doczj.com/doc/242563717.html, //指明访问失败时的联系邮箱 ServerName https://www.doczj.com/doc/242563717.html,:80 //指定服务器域名 DocumentRoot "/opt/ouc-server"
1、安装Apache2 yum install httpd 2、启动 方法一：servicehttpd start 方法二：/etc/init.d/httpd start //浏览http：//ip，应该看到Apache2的测试页 3、设置开机启动 方法一：chkconfig--levels 235 httpd on 方法二：chkconfighttpd on //Apache的默认文档根目录是在CentOS上的/var/www/html 目录，配置文件是/etc/httpd/conf/httpd.conf。 4、安装PHP5 yum install php //重启服务 方法一：servicehttpd restart 方法二：/etc/init.d/httpd restart 测试PHP5是否安装成功 创建info.php vi /var/www/html/info.php 内容：
?> //浏览http：//ip/info.php 5、PHP5获得MySOL的支持 yum search php //还安装需要安装的 yum install php-mysqlphp-gdphp-imapphp-ldapphp-mbstringphp-odbcphp-pear php-xml php-xmlrpc //现在重新启动Apache2的： /etc/init.d/httpd restart //现在刷新http：//ip/info.php，并再次向下滚动到模块部分。现在，你应该找到更多新的模块，包括MySQL模块. 7.安装phpMyAdmin 通过它可以管理你的MySQL数据库。 首先，我们使CentOS系统RPMForge软件库的phpMyAdmin,而不是官方的CentOS 6.2库： 所以需要导入RPMForge的GPG密钥：
Apache服务器配置及安全应用指南 技术创新变革未来
Apache服务安全加固 一.账号设置 以专门的用户帐号和组运行Apache。 根据需要为Apache 创建用户、组 参考配置操作如果没有设置用户和组，则新建用户，并在Apache 配置文件中指定 (1) 创建apache 组：groupadd apache (2) 创建apache 用户并加入apache 组：useradd apache –g apache (3) 将下面两行加入Apache 配置文件httpd.conf中 检查httpd.conf配置文件。检查是否使用非专用账户(如root)运行apache 默认一般符合要求，Linux下默认apache或者nobody用户，Unix默认为daemon用户
Apache服务安全加固 授权设置 严格控制Apache主目录的访问权限，非超级用户不能修改该目录中的内容 Apache 的主目录对应于Apache Server配置文件httpd.conf的Server Root控制项中应为： 判定条件 非超级用户不能修改该目录中的内容 检测操作 尝试修改，看是否能修改 一般为/etc/httpd目录，默认情况下属主为root:root，其它用户不能修改文件，默认一般符合要求 严格设置配置文件和日志文件的权限，防止未授权访问。 chmod600 /etc/httpd/conf/httpd.conf”设置配置文件为属主可读写，其他用户无权限。 使用命令”chmod644 /var/log/httpd/*.log”设置日志文件为属主可读写，其他用户只读权限。 /etc/httpd/conf/httpd.conf默认权限是644，可根据需要修改权限为600。 /var/log/httpd/*.log默认权限为644，默认一般符合要求。
在Linux下配置Apache服务器 ——江湖、孙中霞、李琴一、实验目的 完成本次实训，将能够: ●配置基本的Apache服务器 ●配置个人用户Web站点。 ●配置虚拟目录别名功能。 ●配置主机访问控制。 ●配置用户身份验证功能.。 ●配置基于IP地址的虚拟主机. 二、实验环境 1、RedHat Linux4AS. 2、Apache 2.0 三、实验内容 1.配置基本的Apache服务器 2.配置个人用户Web站点。 3.配置虚拟目录别名功能。 4.配置主机访问控制。 5.配置用户身份验证功能.。 6.配置基于IP地址的虚拟主机。 四、实验要求 在Linux操作系统下配置Apache服务器。 五、注意事项 1.在修配置文件下注意区分大小写、空格。 2.在每次重新开机后都必须启动Apachec服务器。 3.在每次修改完主配置文件后保存起来，必须重启Apachec服务器，如果不重启会 导致配置无效，最终导致实验失败。 六、实验步骤 1、检测是否安装了Apache软件包: A、首先为服务器网卡添加一个固定的IP地址。 B、在Web浏览器的地址栏中输入本机的IP地址，若出现Test Page测试页面（该 网页文件的默认路径为var/www/html/index.html）如下图1所示就说明Apache 已安装并已启动。
另一种方法是使用如下命令查看系统是否已经安装了Apache软件包： [root@rhe14~]# rpm –aq | grep httpd Httpd-suexec-2.0.52-9.ent Httpd-manual-2.0.52-9.ent System-config-httpd-1.3.1-1 Httpd-devel-2.0.52-9.ent 出现以上内容表明了系统已安装Apache软件包。 2、安装Apache软件包 超级用户（root）在图形界面下选择“应用程序”|“系统设置”|“添加/删除应用程序”命令，选择“万维网服务器”软件包组，在单击“更新”按钮就可以安装与Apache相关的软件包。 3、Apache的基本配置 （1）打开终端输入[root@rhe14~]# /etc/rc.d/init.d/httpd start //启动Apache 或者 [root@rhe14~]# apachectl start //启动Apache [root@rhe14~]# apachectl stop //停止Apache服务 [root@rhe14~]# apachectl restart //重启Apache服务 [root@rhe14~]# apachectl configtest //测试Apache服务器配置语法（2）在httpd.conf将Apache的基本配置参数修改、将一些注释的语句取消注释，或将某些不需要的参数注释掉。 （3）将包括index.html在内的相关网页文件复制到指定的Web站点根目下（var/www/html/index.html） （4）重启httpd进程 (5) 在Web浏览器下输入配置的ip地址出现如下图2，那表明基本配置成功了: