United States Patent [19]
`Ewing et al.
`
`[54] SYSTEM FOR READING mE STATUS AND
`FOR CONTROLLING THE POWER
`SUPPLIES OF APPLIANCES CONNECTED
`TO COMPUTER NETWORKS
`
`[76]
`
`Inventors: Carrell W. Ewing. 585 Washington.
`Palo Alto, Calif. 94301; Andrew J.
`Cleveland. 12467 Creek View Ct.. San
`Martin. Calif. 95046
`
`[21] Appl. No.: 08/685.436
`
`Jul. 23, 1996
`
`[22] Filed:
`Int. Cl.6
`•••••••••••••••••••.••••••••••• G06F 15/16; G06F 1/26
`(51]
`[52] U.S. Cl ................................. 395/200.32; 395/200.53;
`3951750.08
`[58] Field of Search ..................................... 340/644. 646.
`340/310.01. 825.5; 395/200.32. 200.53.
`200.54
`
`[56]
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`4,101,878
`4.495.568
`4,709,318
`4.780,714
`4,937,561
`5,374,922
`5,412.645
`5,481,730
`5,485.576
`5.495,607
`5.506.573
`5.561,769
`
`7/1978 Shimizu eta! .................... 340/310.QI
`111985 Gilbert eta! ........................ 395/182.2
`1111987 Gephart et a! ............................ 363/37
`10/1988 Moustakas et al .................. 340/825.5
`6/1990 Sasaki eta! ............................ 340/646
`12/1994 Ebersohl ................................. 340/644
`511995 Younkin et al ......................... 370/365
`111996 Brown eta! ....................... 395n50.08
`1/1996 Fee et al ........................... : 395/185.09
`2/1996 Pisello et a!. ............................. 707/10
`4/1996 Ewing et al.
`........................... 340/644
`10/1996 Kumar eta! ....................... 3951200.32
`
`I IIIII
`
`11111111111111111 Will IIIII IIIII Wllllllllllllllllllllllll
`5,949,974
`Sep.7,1999
`
`US005949974A
`[11] Patent Number:
`[45] Date of Patent:
`
`5,585,678
`5.652,893
`5.761.084
`5,781,434
`
`1211996 Dijk et al ................................ 3071112
`7/1997 Ben-Meir et al .................. 395n50.02
`6/1998 Edwards ............................. 364/528.28
`7/1998 Tobita et al ............................. 364/184
`
`OTHER PUBLICATIONS
`
`W. Richard Stevens. ''TCPIIP Illustrated. vol. 1 -The
`Protocols". pp. 359-361. 1994.
`Michael Slater. "Microprocessor-Based Design-A Com(cid:173)
`prehensive Guide to Hardware Design". pp. 19-24. 1989.
`Peter Drake. "Using SNMP to Manage Networks", pp.
`2/1-2/4. 1991.
`
`Primary Examiner-Zarni Maung
`Assistant E=miner-Andrew Caldwell
`Attome); Agent, or Firm-Richard B. Main
`ABSTRACT
`
`[57]
`
`A SNMP network comprises a power manager with a SNMP
`agent in TCPIIP communication over a network with a
`SNMP network manager. The power manager is connected
`to control several intelligent power modules each able to
`independently control the on/off power to several inter(cid:173)
`networking devices in an equipment rack at a common
`remote node. e.g .. a point-of-presence site. Power-on and
`load sensors within each intelligent power module are able
`to report the power status of each inter-networking device to
`the SNMP network manager with Mffi variables in response
`to GEf commands. The SNMP network manager is further
`able to reboot each inter-networking device by cycling the
`power on/off to its respective intelligent power module with
`the SEf command provided in conventional SNMP man(cid:173)
`agement applications.
`
`12 Claims, 4 Drawing Sheets
`
`10~
`
`14
`
`TCP/IP
`
`r;.
`50~
`~ 48
`
`12
`
`host
`
`26
`
`utility power
`
`IPR Page 1
`
`Raritan v. Server Technology
`
`RARITAN EXHIBIT 1012
`
`

`
`~
`....J
`\0
`\0
`~
`\0
`01
`
`'41
`
`'41
`
`.a:;..
`~
`0
`
`\FJ =-a .....
`
`~
`
`1-"
`.;'!
`~ -'!='
`
`\J).
`
`~ = ......
`
`~ ......
`~
`•
`00
`
`e •
`
`utility power
`
`L-~-------,--------~ 26
`
`UPS
`
`30
`
`32
`
`Fig.~1
`
`34
`I
`
`~
`
`1
`
`agent 1 1
`SNMP
`
`liP
`
`I
`
`I
`
`I I
`
`I power 1 46
`
`mgr
`
`28
`
`1}}
`
`POP ,....----'---.
`
`agent
`SNMP
`
`I 12
`
`20
`
`I~
`
`host
`
`manager
`SNMP
`
`TCP/IP
`
`14
`
`10~
`
`~48
`.-s.::::t~_
`
`5~
`
`IPR Page 2
`
`

`
`U.S. Patent
`
`Sep.7, 1999
`
`Sheet 2 of 4
`
`5,949,974
`
`Fig._2
`
`100-.
`
`102
`
`apply a series of alternating
`current (AC) voltage pulses
`to an appliance with an on/off
`switch that are synchronized
`to a source of AC power
`
`~,
`sense the presence of any series of
`AC current pulses that result
`if the appliance switch is closed
`
`r-104
`
`~'
`analyze any AC current pulses detected
`in step 1 04 to determine if
`they resulted from the application
`of the AC voltage in step 1 02
`
`1
`
`106
`
`output an on/off status indication
`for the appliance switch
`
`1 08
`
`~,
`
`IPR Page 3
`
`

`
`~
`"' \0
`\0
`.a;.
`"' \0
`Ul
`
`Fig._3
`
`serial I/O
`
`---........ ----..... -........ --------......... -----................. ---................ -.... ------.. :-_-_-_-_-_-_-_-_-_-_-. .-.. :-... · .. ::! __ ......... ----------------6-----.... ---................ ..
`
`221
`
`'
`i
`'
`! ~-----.-___J
`
`I
`
`0
`
`.. i
`
`'
`
`I
`
`R
`
`0
`
`-+---1 ./
`........
`242 !
`, ...... ®. -----EACG COMM
`non/off V-SENS
`
`('!) -w
`
`rJ:J =- ('!)
`
`,I;;.
`"""=
`0
`
`~
`
`\C)
`
`~
`
`~-....1
`'F'
`('!)
`rJ:J
`
`~ ;-= """"
`00. •
`c: •
`
`224
`
`load rT
`device
`networl
`
`1 . l_oa~ .. ~~?
`,
`
`. ,
`
`0
`I
`I
`
`I
`I
`I
`
`214
`
`:i-....,
`
`238
`
`11:
`i
`:
`:
`!
`!
`:
`:
`!
`i
`!
`!
`:
`:
`i i
`i
`i
`
`----------·---------~
`
`,1,H ~--<
`
`0
`0
`
`0
`I
`0
`0
`
`I
`I
`
`:
`
`/~ f23Q ~rs
`
`0
`
`I
`
`236
`
`:
`1 •••••••••••••••••••••• ~~~-~----·····-···········-··-··············-····························································l
`
`I
`
`: ~
`p
`:
`I ;
`*' .,.. _,_ T
`• .!.~ l
`I)
`!211
`HL Y
`:········· ------··-·-······· --~
`
`L ........ ···-········-----····j
`:
`,--~
`
`I
`
`I
`
`•
`
`I
`
`I
`
`I
`
`'
`
`70VAC
`
`.....,..
`
`1---
`
`GND
`
`N
`
`Clock
`Supply + 15V
`Power
`
`&
`
`Generator +5V
`
`CLK
`
`r-212
`
`H
`
`:
`n
`~ c
`~ :::-
`st ~
`0 ~
`Q 15 '225
`i
`
`l N
`
`o
`\1 0
`..... 1
`
`U>:
`
`(")
`
`CD
`
`0 3
`C/) 0
`
`--4
`
`:
`
`0
`I
`I
`I
`I
`I
`
`0 ..
`
`0
`
`38, 40, 42, 4
`4
`
`: ........... , .... .
`:
`
`)
`
`240
`
`J.
`)..G
`0 °
`:
`:
`
`I
`
`I
`I __ , __ 0
`
`CLK 1-SENS
`I
`+
`
`5V
`
`I
`220,
`
`I
`
`I I
`
`IPR Page 4
`
`

`
`U.S. Patent
`
`Sep.7, 1999
`
`Sheet 4 of 4
`
`5,949,974
`
`Fig·-4
`214
`,------------------------------------(_ __________________________________________________________ ------
`
`+5 VDC
`
`70VAC
`
`246
`
`+12 voc
`
`power
`supply
`
`Eref
`
`244
`
`Rs
`
`250
`
`I_ SEN
`
`'
`'
`·-~ .... · - - - - - - - ..... -J
`
`.
`.
`
`236
`
`hot
`
`Fig._5
`
`+5VDC
`216
`---------- --------_/._ ___________ --
`
`256
`________ /.__~_~2
`
`254
`
`V _SENS ._----:-----i--1
`
`neutral
`
`238
`
`IPR Page 5
`
`

`
`5.949.974
`
`1
`SYSTEM FOR READING THE STATUS AND
`FOR CONTROLLING THE POWER
`SUPPLIES OF APPLIANCES CONNECTED
`TO COMPUTER NETWORKS
`
`BACKGROUND OF THE INVENTION
`1. Field of the Invention
`The invention relates generally to automatic power con(cid:173)
`trol and more particularly to remote control methods and
`devices to maintain computer network system availability.
`2. Description of the Prior Art
`Enterprise networks exist to support large world-wide
`organizations and depend on a combination of technologies.
`e.g .. data communications. inter-networking equipment
`(frame relay controllers. asynchronous transfer mode (ATM)
`switches, routers. integrated services digital network (ISDN)
`controllers. application servers). and network management
`application software. Such enterprise networks can be used
`to support a large company's branch offices throughout the
`world. and, as such. these networks have become mission
`critical to the functioning of such organizations. Masses of
`information are routinely expected to be exchanged. and
`such information exchanges are necessary to carry on the
`daily business of modern organizations. For example. some
`international banks have thousands of branch offices placed
`throughout Europe. Asia and the United States that each
`critically depend on their ability to communicate banking
`transactions quickly and efficiently with one another and
`headquarters.
`A typical enterprise network uses building blocks of
`router and frame relay inter-networking devices mounted in
`equipment racks. Such equipment racks are distributed to
`remote point of presence (POP) locations in the particular
`network. Each equipment rack can include frame relay 35
`controllers. routers. ISDN controllers. servers and modems.
`etc., each of which are connected to one or more power
`sources. The value of POP equipment can range from
`$200.000 to $500.000. and the number of individual devices
`can exceed a thousand.
`Many enterprises rely on an uninterruptable power supply
`(UPS) to keep their inter-networking devices operational.
`Many inter-networking devices are typically connected to a
`single UPS. and this sets up a problem. When an individual
`router locks up. the router's power cannot be individually
`cycled on and off externally at the UPS because it is
`connected to a multiple power outlet. The recovery action
`choices available to the network control center operator thus
`do not include being able to reinitialize the individual
`equipment through a power interruption reset. The network
`operator could command the UPS to power cycle. but that
`would reset all the other attached devices that were osten(cid:173)
`sibly operating normally and carrying other network traffic.
`Another option is to dispatch someone to the remote location
`to reset the locked-up device. Neither choice is an attractive 55
`solution.
`In large organizations that have come to depend heavily
`on enterprise networks. great pressures develop to control
`costs and thus to improve profits. Organizational down(cid:173)
`sizing has been used throughout the corporate world to 60
`reduce non-network costs. and that usually translates to
`fewer technical people available in the right places to
`support large and complex in-house global networks. Such
`reduced repair staffs now rely on a combination of central(cid:173)
`ized network management tools and third-party maintenance 65
`organizations to service their remote POP sites. The costs
`associated with dispatching third-party maintenance techni-
`
`50
`
`2
`cians is very high. and the dispatch and travel delay times
`can humble the business operations over a wide area for
`what seems an eternity.
`Global communication network operators. located at a
`5 few centralized network management centers. are relying
`more and more on automated network management appli(cid:173)
`cations to analyze. process. display and support their net(cid:173)
`works. An increasing number of network management soft(cid:173)
`ware applications are being marketed that use open-system
`10 standardized protocols. Particular network application tool
`software is available to report lists of the inter-networking
`devices. by location. and can issue trouble lists and keep
`track of software versions and releases. New sin1ple network
`management protocol (SNMP) applications are convention-
`15 ally used to issue alarms to central management consoles
`when remote inter-networking devices fail.
`One such SNMP network management application is
`marketed by Hewlett-Packard. HP OPENVIEW is a family
`of network and system management tools and services for
`20 local and wide area multi vendor networks. OPENVIEW is a
`management platform that provides application developers
`and users with the ability to manage multivendor networks
`and expand their distributed computing environments.
`OPENVIEW allows network operation centers to build an
`25 intelligent hierarchical network management application.
`and uses open standards such as SNMP. user datagram
`protocol (UDP). and tlle now ubiquitous transmission con(cid:173)
`trol protocoVinternet protocol (TCP/IP). Because OPEN(cid:173)
`VIEW is built on open system standards. global communi-
`30 cation network operators can easily integrate the various
`inter-networking equipment nodes into a managed environ(cid:173)
`ment operated by strategically located network consoles.
`In order to provide a reliable computing environment. a
`robust and active process for problem resolution must be in
`place. OPENVIEW allows the definition of thresholds and
`monitoring intervals. and the interception of network.
`system. database. and application-messages and alerts. Once
`a threshold value is. exceeded. intelligent agents can run a
`40 pre-defined automatic action and/or generate and send a
`message to alert an operator on a central management
`console. Messages can also be forwarded to a pager or
`trouble-ticketing application. To help focus on the most
`critical problems. a message browser window is used to
`45 display six severity levels for incoming problems and
`events. e.g .. ranging from stable to critical. An integrated
`history database is provided for auditing and analyzing
`system and network activities. for identifying trends and for
`anticipating problems before they occur. Activity displays
`and reports can be customized by the users.
`Prior art SNMP network management uses embedded
`microprocessors in almost every inter-networking device to
`support two-way inter-computer communications with TCP/
`IP. of which SNMP is a member of the TCPIIP protocol
`suite. SNMP is conventionally used to send messages
`between management client nodes and agent nodes. Man(cid:173)
`agement information blocks (Mffis) are used for statistic
`counters. port status. and other information about routers
`and other network devices. GET and SET commands are
`issued from management consoles and operate on particular
`Mffi variables for the equipment nodes. Such commands
`allow network management functions to be carried out
`between client equipment nodes and management agent
`nodes.
`SNMP is an application protocol for network manage(cid:173)
`ment services in the internet protocol suite. SNMP has been
`adopted by numerous network equipment vendors as their
`
`IPR Page 6
`
`

`
`5.949.974
`
`3
`main or secondary management interface. SNMP defines a
`client/server relationship. wherein the client program. a
`"network manager". makes virtual connections to a server
`program. an "SNMP agent". on a remote network device.
`The data base controlled by the SNMP agent is the SNMP
`management information base. and is a standard set of
`statistical and control values. SNMP and private MJBs allow
`the extension of standard values with values specific to a
`particular agent. Directives issued by the network manager
`client to an SNMP agent comprise SNMP variable 10
`identifiers, e.g .. MJB object identifiers or MJB variables. and
`instructions to either GET the value for the identifier. or SET
`the identifier to a new value. Thus private MIB variables
`allow SNMP agents to be customized for specific devices,
`e.g .. network bridges, gateways. and routers. The definitions 15
`of MJB variables being supported by particular agents are
`located in descriptor files. typically written in abstract syn(cid:173)
`tax notation (ASN.l) format. The definitions are available to
`netw()rk management client programs.
`SNMP enjoys widespread popularity. and SNMP agents 20
`are available for network devices including computers.
`bridges. modems. and printers. Such universal support pro(cid:173)
`motes interoperability. The SNMP management protocol is
`flexible and extensible. SNMP agents can incorporate device
`specific data. Mechanisms such as ASN.l files allow the 25
`upgrading of network management client programs to inter(cid:173)
`face with special agent capabilities. Thus SNMP can take on
`numerous jobs specific to device classes such as printers.
`routers. and bridges. A standard mechanism of network
`control and monitoring is thus possible.
`Unfortunately. SNMP is a complicated protocol to
`implement. due to complex encoding rules, and it is not a
`particularly efficient protocol. Bandwidth is often wasted
`with needless information. such as the SNMP version that is
`to be transmitted in every SNMP message. and multiple 35
`length and data descriptors scattered throughout each mes(cid:173)
`sage. SNMP variables are identified as byte strings. where
`each byte corresponds to a particular node in the MIB
`database. Such identification leads to needlessly large data
`handles that can consume substantial parts of each SNMP 40
`message.
`Most vendors implement network managers thinking a
`user's primary interest is in the data associated with par(cid:173)
`ticular network devices. But such data is easily acquired by 45
`other means. e.g .. "netstat" and "rsh" UNIX programs. The
`important information about the network includes the dif(cid:173)
`ferences between devices. besides their current states.
`SNMP affords a good mechanism for rapidly processing
`such differences on large networks. since SNMP avoids the 50
`processing burden of remote login and execution.
`Network management applications can thus monitor the
`health of every part of a global communications network and
`can be set to communicate alarms to a central management
`console. Current network management applications do an 55
`adequate job of informing central management consoles
`about the health of various nodes in the network and the
`alarms they issue when a node is failing are useful.
`Conventional SNMP network management technologies
`do not provide sufficient information related to the nodes' 60
`electrical power status. A new technology is needed that can
`be simply and inexpensively added to client equipment
`nodes for SNMP reporting of the electrical power status of
`the node. For example. in a router based network with
`SNMP support. prior art individual routers can use SNMP to 65
`issue an alarm to the management console. But the console
`operator would know only that the router is failing. A GET
`
`4
`command can be issued to the router node to determine if the
`counter and buffer thresholds limits were exceeded and
`caused a router to lock-up. However. the console operator
`does not have any information about the electrical power
`status to the router. e.g .. has the router power switch been
`moved to the OFF position or has the switch been acciden(cid:173)
`tally turned OFF? The electrical power source could have
`failed. the power cable connection become loose. or a
`technician may have accidentally removed the router from a
`rack.
`
`SUMMARY OF THE PRESENT INVENTION
`
`It is therefore an object of the present invention to provide
`a system and method for providing power supply status and
`control in network nodes at geographically distant locations.
`It is another object of the present invention to provide a
`system and method for describing power supply status and
`control in SNMP MIB variables between network nodes and
`a central network management console.
`Briefly. an SNMP network embodiment of the present
`invention comprises a power manager with an SNMP agent
`in TCPIIP communication over a network with an SNMP
`network manager. The power manager is connected to
`control several intelligent power modules each able to
`independently control the power on/off status of several
`inter-networking devices in an equipment rack at a common
`remote node. e.g .• a point-of-presence site. Power-on and
`load sensors within each intelligent power module are able
`30 to report the power status of each inter-networking device to
`the SNMP network manager with MJB variables in response
`to GET commands. The SNMP network manager is further
`able to reboot each inter-networking device by cycling the
`power off/on to its respective intelligent power module with
`the SET command provided in conventional SNMP man(cid:173)
`agement applications.
`An advantage of the present invention is that a system and
`method is provided for communicating and controlling the
`power supply status of network nodes at geographically
`distant locations.
`Another advantage of the present invention is that a
`system and method is provided for describing power supply
`status and control in SNMP MJB variables between network
`nodes and a central network management console.
`A further advantage of the present invention is that a
`system and method is provided that allows a network
`console operator to investigate the functionality of the
`electrical power status when a router or other network
`device has been detected as failing.
`A still further advantage of the present invention is that a
`system and method is provided for reducing the need for
`enterprise network operators to dispatch third party main(cid:173)
`tenance vendors to remote equipment rooms and POP loca(cid:173)
`tions simply to power-cycle failed inter-networking devices.
`The costs to dispatch such third party maintenance vendor
`can run from $300-$600 per call. The cost of implementing
`the present invention can be recaptured in less than one year.
`e.g .. by reducing the number of third party maintenance
`dispatches to remote locations.
`Another advantage of the present invention is that a
`system and method is provided for reducing the time it takes
`to restore a failed inter-networking device and improving
`service level measures.
`Another advantage of the present invention is that a
`system and method is provided for reducing organization
`losses from network downtime. Being able to inunediately
`
`IPR Page 7
`
`

`
`5.949.974
`
`5
`power-cycle a failed server and thus return the server to
`operation can directly reduce the downtime loss to the
`organization.
`These and many other objects and advantages of the
`present invention will no doubt become obvious to those of 5
`ordinary skill in the art after having read the following
`detailed description of the preferred embodiments which are
`illustrated in the various drawing figures.
`
`IN THE DRAWINGS
`FIG. 1 is a block diagram of a simple network manage(cid:173)
`ment protocol (SNMP) network embodiment of the present
`invention:
`FIG. 2 is a flowchart of a method of appliance power
`switch status detection. according to the present 'invention:
`FIG. 3 is a schematic of a representative intelligent power
`module such as are included in the network of FIG. 1:
`FIG. 4 is a schematic diagram of the load sensor included
`in the intelligent power module of FIG. 3; and
`FIG. 5 is a schematic diagram of the power-on sensor
`included in the intelligent power module of FIG. 3.
`
`25
`
`30
`
`6
`FIG. 3 illustrates an intelligent power module 200, similar
`to intelligent power modules 30, 32. 34. 36. which may be
`located external or internal to devices 38. 40, 42. 44. or
`internal or external to the UPS 26. The intelligent power
`module 200 includes a power supply and clock generator
`212. a load sensor 214. a power-on sensor 216. a solid-state
`relay 218 and a microprocessor 220. A serial input/output
`(110) connection 221 provides for communication with a
`controller. e.g .. power manager 28. An appliance. such as the
`10 inter-networking devices 38. 40. 42. 44. has a power on/off
`switch 223 that may be internal or external to the appliance
`represented by a network device load 224. The switch 223
`may also actually comprise both internal and external
`switches in series. The incoming alternating current (AC)
`15 line power is applied to the intelligent power module 200 at
`a hot (H) terminal 225. a neutral (N) terminal 226 and a
`ground (G) terminal228. The appliance has its incoming AC
`line power applied to a hot (H) terminal 230. a neutral (N)
`terminal 232 and a ground (G) terminal 234. which are
`20 respectively connected to a hot (H) terminal 236. a neutral
`(N) terminal238 and a ground (G) terrninal240. A relay 242
`allows automatic remote control by the microprocessor of
`power to the appliance due to its position in the incoming
`AC line.
`The load sensor 214 is such that if a current is flowing
`because switch 223 is closed. the microprocessor will
`receive a logic low status tltdication.
`FIG. 4 represents an embodiment of the load sensor 214
`included in FIG. 3. The load sensor 214 comprises a sense
`resistor 244 connected to a voltage comparator 245. When
`the voltage dropped across the sense resistor 244 exceeds a
`reference voltage provided by a power supply 246. the
`output of the voltage comparator 245 goes high. A resistor
`247 couples this to an opto-isolator 248 and produces a five
`35 volt digital output (I_SENS) that indicates load/no-load to
`the microprocessor 220.
`FIG. 5 represents an embodiment of the power-on sensor
`216 included in FIG. 3. The power-on sensor 216 includes
`40 an opto-isolator 252. The output of the opto-isolator 252
`goes low when a sufficient voltage is dropped across a
`resistor 254. A five volt power supply connection and a
`pull-up 256 provide a five volt logic output (V _SENS) that
`indicates power/no-power to the microprocessor 220.
`In operation, the device 200 senses if switch 223 is closed
`or open by converting AC current pulses from the power
`supply 212 that flow through the series circuit comprising
`the solid-state relay 218. the H-terminals 236 and 230. the
`switch 223. the network device load, the N-terrninals 232
`50 and 238. the load sensor 214. and return to the power supply
`212. If the switch 223 is open. no such current can flow.
`The power supply and clock generator 212 provides a five
`volt pulse clock (CLK) to the microprocessor 220 at each
`zero-crossing of the incoming AC power line voltage across
`55 the H-terminal 225 and the N-terminal 226. A slightly
`delayed version of the clock is output by the microprocessor
`220 to control the solid-state relay 218. A seventy volt AC
`output (70VAC) of the power supply and clock generator
`212 provides a reduced voltage AC sine wave that is
`60 approximately seventy volts RMS. The solid-state relay 218
`therefore gates through the seventy volt AC waveform twice
`each cycle such that alternating pulses of +70 volts and -70
`volts are sent through switch 223 and load sensor 214. If a
`current flows because the switch 223 is closed. a character-
`65 istic pulse synchronized to the CLK signal will appear as an
`output from the opto-isolator 248. A resistor 250 provides a
`pull-up to a current sense input to the microprocessor 220.
`
`DEf~ED DESCRIPfiON OF THE
`PREFERRED EMBODIMENTS
`FIG. 1 illustrates a simple network management protocol
`(SNMP) network embodiment of the present invention.
`referred to herein by the general reference numeral10. The
`SNMP network 10 includes a host 12 with a TCPIIP con(cid:173)
`nection 14 to a plurality of point-of-presence (POP) nodes
`represented by a pair of equipment racks 16 and 18. SNMP
`network management is provided by a SNMP manager 20 in
`communication with a respective pair of SNMP agents 22
`and 24 at the remote nodes. The SNMP manager 20 may
`comprise a commercial product such as ffiM NEfVIEW/
`6000. HP OPENVIEW, POLYCENTER. SunNet
`MANAGER. Cabletron SPECTRUM. etc.
`Each POP node can be associated with an uninterruptable
`power supply (UPS) 26 that provides operating power to a
`TCPIIP-addressable enterprise power manager 28 and sev(cid:173)
`eral intelligent power modules 30. 32. 34. 36 connected to
`corresponding inter-networking devices 38. 40. 42, 44. An
`SNMP agent 46 is private to the power manager 28 and does
`not depend· on the equipment rack 16 or its inter-networking
`devices 38, 40. 42. 44. The power manager 28 is connected 45
`to independently control each of the intelligent power mod(cid:173)
`ules 30. 32, 34, 36. Such control includes being able to sense
`the power-on and load status of each of the inter-networking
`devices 38, 40, 42, 44 and to switch power on and off to each
`of the inter-networking devices 38. 40. 42. 44. Such status
`is sensed and reported by an SNMP GEf command 48 and
`the power switching is accomplished with an SNMP SEf
`command 50 that issue from the host 12.
`FIG. 2 shows a method of appliance power switch status
`detection. referred to herein by the general reference
`numeral100. The method 100 comprises a step 102 applying
`a series of alternating current (AC) voltage pulses to an
`appliance with an on/off switch that are synchronized to a
`source of AC power. A step 104 senses the presence of any
`series of AC current pulses that result if the appliance switch
`is closed. A step 106 analyzes any AC current pulses
`detected in step 104 to determine if they resulted from the
`application oftheAC voltage in step 102. A step 108 outputs
`an on/off status indication for the appliance switch. Method
`100 does not result in the turning-on and the operation of the
`appliance during steps 102 or 104. and is therefore unob-(cid:173)
`trusive.
`
`IPR Page 8
`
`

`
`5.949.974
`
`10
`
`7
`If the switch 223 is open. the characteristic pulses will not
`appear. An "onsense" device 252 provides isolation for a
`voltage sense input to the microprocessor 220.
`The microprocessor 220 analyzes and stores its determi(cid:173)
`nation of whether the power is applied to the device 38-44 5
`and whether the switch 223 is closed. Such data is thereafter
`useful to control the relay 242. The microprocessor 220 is
`programmed to control the relay 242 and to report the
`presence of current and voltage to the appliance through
`serial communication conducted over the serial 110 connec-
`tion 221.
`The power manager 28 is able to read from the intelligent
`power modules 30, 32. 34. 36. whether there is a proper
`operating voltage being supplied to the inter-networking
`devices 38. 40. 42. 44. and whether such loads are turned on.
`The power manager 28 and its SNMP agent 46 are able to
`report such status in response to the GEf command 48. The
`GEf command modifies a M1B variable that is reported by
`the SNMP agent 46 to the SNMP manager 20.
`The power manager 28 is able to require the intelligent 20
`power modules 30. 32. 34. 36. to turn the power being
`supplied to the inter-networking devices 38. 40. 42. 44. on
`or off in response to the SEf command 50. Such SEf
`commands modify the MIB variable defined for power
`on/off. and allow independent power-cycling of each and 25
`any of the inter-networking devices 38, 40, 42. 44. Such
`power cycling promotes a power-up reset of the appliance,
`e.g .. when the SNMP agent 22 has reported a failure of the
`POP node 16 to the SNMP manager 20.
`SNMP defines a client/server relationship. The client 30
`program. network manager 20, makes virtual connections to
`the server program. the SNMP agent 22 and 24 on a remote
`network device. The database controlled by the SNMP agent
`is the management information base (MIB). The M1B is a
`standard set of statistical and control values that provides 35
`information about the attributes of devices attached to the
`network. SNMP allows for the extension of these standard
`values with values that are specific to a particular SNMP
`agent through the use of private MIBs. The use of private
`MIB variables allows SNMP agents to be modified for a 40
`variety of devices, e.g .. bridges. hubs. routers and CSU/
`DSUs. etc. SNMP operates by exchanging network infor(cid:173)
`mation through protocol data unit (PDU) messages. PDUs
`carry variables that have both titles and values. There are
`five types of PDUs that SNMP uses to monitor a network, 45
`two for reading terminal data, two for setting terminal data.
`and one. the trap. monitoring network events. Every SNMP
`message consists of a variable, and every variable consists
`of a variable title. the integer. string data type of the variable.
`whether the variable is read-only or read-write. and the value so
`of the variable.
`The SNMP manager 20 collects information via MIBs
`about routers, hubs, bridges. concentrators, servers. switches
`and other inter-networking devices. When a problem at a
`remote node is detected. the corresponding SNMP agent
`issues an alarm that identifies the problem by type and node
`address. The SNMP manager typically sends a Telnet script
`to a TCPJIP-addressable enterprise power manager. The
`Telnet script instructs the enterprise power manager to cycle
`the power cycle. to recover an otherwise locked-up network
`device. SNMP management is not required for the enterprise
`power manger and the associated intelligent power modules.
`The intelligent power modules include normally closed
`relays so power is always on except when the relay is
`deliberately opened to trigger a power on reset and reboot.
`The network management application monitors the UPS and
`the inter-networking devices.
`
`8
`The load sensor and power-on sensor can be combined
`such that a console operator can determine if electrical
`power is available to an equipment rack and to an individual
`inter-networking device. A relay reset located between the
`power source and the client equipment node supports an
`SNMP-type SET command that can be defined to open and
`close a relay to power-cycle the inter-networking device.
`Such power-cycling can clear a lockup condition and allow
`the device to return to normal operation via its own internal
`power-up reset mechanism.
`A console operator can be notified by conventional means
`that a router is failing.