Patent Application
20090097200
The present invention relates generally to telecommunications, networking and computer equipment and specifically, but not exclusively, to a modular Advanced Telecom Computing Architecture based server blade arrangement that is interoperable and interchangeable with conventional server blades in the enterprise setting and that is configurable into an interconnected structure capable of providing scalable mechanical, electrical and environmental functionality.
One of the products that has deep penetration into the enterprise information technology infrastructure is the ubiquitous multi-tiered modular rack supporting a group of modular chassis that in turn are capable of meeting, either singly or in combination with other chasses, the data communication, computing and telecommunication needs of the enterprise. Typically, racks provide each chassis access to network connections and electrical power. A typical chassis may contain components and interconnecting devices such as printed circuit boards, interconnecting wires, electronic and mechanical modules and other components arranged to form a system such as a server. Each chassis functions to provide an environment relatively free from excessive heat, shock, vibration and/or dust for the computer system. Groups of these computer systems are interconnected to form electronic applications, such as server farms that serve the networking needs of business organizations. Prior art practice has found it advantageous to standardize rack dimensions in order to house standardized chasses. For example, a 19-inch rack is standardized (EIA 310-D, IEC 60297 and DIN 41494 SC48D) for mounting various electronic modules in a “stack”, or rack, 19 inches (482.6 mm) wide. Chasses designed to be placed in a rack are designated as rack-mount, a rack mounted system, a rack mount chassis, subrack, or occasionally, simply shelf. A 1 U tall subrack (also referred to as a “pizza box” or “slab”) chassis is available in various form factors such as 19 inches wide by 24, 30, 36 and 48 inches deep. A “U” is the measured height of a chassis case where 1 U=1.75 inches or 44.45 mm. 19 inches or 482.6 mm is the width of the front plate, 17.75 inches or 450.85 mm the space left for the chassis (the rails take up 0.625 inches or 15.875 mm on each side). Each chassis may house a separate computer or server, for example, having one or more CPUs. In particular, the small form factor allows for a large number of servers to be vertically stacked, typically with up to 47 U in each rack. The modular nature of each such chassis allows for a given server system to be swapped out of the network and the rack without interfering with the operation of other computer systems.
Conventional server blades while conforming to some industry specification in terms of the external form factor (geometry) so they can be mounted on a standardized rack or in terms of the input/output (I/O), power and other interfaces they present to the environment external to the blade so as to be compatible for cooperative operation with the rack and other external devices on and off the rack. Generally, such blades embody proprietary, monolithic electro-mechanical dedicated hardware and software solution within that external form factor that are designed for a particular application or set of applications and made available in a packaging that typically conforms to an industry-wide standard. The internal geometry and layout is relatively fixed in that they cannot be easily configured to suit the needs of an alternate application except perhaps by swapping out the blade with a blade specifically designed for the alternate application. Conventional blade architecture generally provides both, the electronic circuitry—i.e. devices, components, firmware and software, that actually effectuate a desired application, as well as the auxiliary modules/subsystems such as the cooling units, the controllers, the management software and firmware and so forth that play a supporting role in effectuating the desired application. Changing out a conventional server module does not generally allow the retention and reuse of the auxiliary modules with the replacement blade. Nor can the dedicated hardware solution available in a particular sized chassis be easily adapted to a different sized chassis targeted for operation in a different server environment. Very often the replacement blade is from the same manufacturer or vendor when several blades need to operate cooperatively because of the difficulties of integrating blades from different vendors in the same rack or enterprise. Accordingly, it would be advantageous to provide a blade, targeted to the enterprise class server technology segment, that can be constructed from commercial-off-the-shelf (COTS) technology and that can be swapped out, if needed, for another blade also based on COTS technology but capable of delivering a different application(s) without having to discard or replace the auxiliary modules/subsystems that are oftentimes associated with and integrated into the blade.
There has been a widespread shift from the historic telecommunications business model which fostered low unit volume, relatively high price proprietary system architectures to standards-based solutions built using COTS technology. One of the business drivers for this shift is the need for flexibility to respond to a rapidly changing network infrastructure and the need to keep operating and capital expenditures low. Catalyzing this shift are standards based technologies that adhere to specifications defined by industry sponsored standards making bodies. For example, the Advanced Telecom Computing Architecture (or AdvancedTCA™ hereinafter “ATCA”) based platform can be used by both, suppliers and end-users to construct ATCA standard-compliant solutions.
The ATCA specification is a series of industry standards that define scalable, standardized platform architecture to extend COTS to a broad spectrum of products available from component vendors. ATCA compliant components and systems embody interoperable ATCA technology such as physical format, system management and software designed to deliver cost effective, reduced time-to-market, off-the-shelf solutions that can be incorporated into products ranging from high-availability, carrier-grade telecom, storage, and computing applications. ATCA is sponsored by the PCI (Personal Computer Interconnect)—Industrial Computer Manufacturers Group (PICMG®), a major industry standards body.
The ATCA Base Specification, PIGMG 3.0 Revision 2.0, ratified in Dec. 30, 2002 (hereinafter “the ATCA specification”), defines an open electromechanical architecture of a modular platform that may be constructed from commercial off-the-shelf components. The electromechanical architecture encompasses the rack and shelf (chassis) mechanical form factors, power parameters, cooling characteristics, core backplane fabric interconnects and system management architecture to enable the construction of a modular platform that is capable of receiving a multiplicity of ATCA compliant modular plug-in circuit boards (ATCA Carrier cards). The ATCA compliant modular plug-in circuit boards feature an open electromechanical architecture also defined by the ATCA specification. The ATCA base specification together with other associated specifications define multiple fabric connections and support multiple protocols for control and data plane communications including Ethernet, Fibre Channel, InfiniBand, StarFabric, PCI Express, and RapidIO®.
The AdvancedTCA specification defines the requirements for the ATCA circuit boards when plugged into ATCA circuit board slots provided in the ATCA backplane to form an ATCA system or shelf. The ATCA 3.0 base specification defines a power budget of 200 Watts (W) per board. Power is delivered to the frame by dual redundant −48 VDC feeds. Local DC-DC conversion is accomplished per board. Redundant local power feeds are normally attached through either diode OR'ed connections to a single on-board DC-DC converter or via on-board dual redundant load sharing DC-DC converters. The ATCA backplane includes power distribution circuitry that distributes the dual redundant power feeds from the two power entry modules to the circuit boards plugged into the ATCA backplane such that as much as 200 Watts is dissipated per single-slot ATCA circuit board in addition to the power consumption requirements of other ATCA-specific shelf components. Circuit boards may occupy more than one circuit board slot to receive more than the maximum 200 Watts (W) per board. Under the PICMG 3.0 specification, each board is supplied with dual −48 VDC feeds. Each feed remains isolated and is fed individually to each board slot through the backplane. There are two basic methods for combining the dual redundant feeds. One method combines the two feeds through diode OR'ing and delivers the combined single feed to DC-DC converters. If either feed fails, all power shall be delivered by the surviving feed. The second method is to direct each of the two feeds to its own DC-DC converter. The outputs of the converters are then combined to provide power to the rest of the power supply circuits.
In order to support power dissipation of 200 W per board slot, there needs to be a mechanism for adequate cooling to prevent overheating and resulting failure of the devices and components of the system. The AdvancedTCA specification prescribes the rate of cooling air flow through partitions of the AdvancedTCA board and shelf sufficient to dissipate 200 W of power per board slot. In a typical implementation, blowers (i.e. mechanical fans) are provided in the shelf to pull air from front to rear and bottom to top. Thermal designs using mechanical fans are also typically encountered in enterprise class, rack mounted server blades. The standard cooling configuration in conventional rack mounted server blades is generally from front to rear and bottom to top in the manner specified for the AdvancedTCA shelf. In some instances, the rack in which the blades are housed is also tasked with generating and supplying the air flow.
Each of these approaches to cooling has several drawbacks. Notably, it assumes that cooling requirements for all boards are substantially equal, and that the airflow across all the boards is approximately equal. Furthermore, it does not consider “hot spots” on individual boards, but rather again uses an average airflow approach. In practice, the power consumption (and thus heat generation) within a typical ATCA chassis is uneven, with certain types of boards producing more heat that other types of boards. Furthermore, in many boards only a few components, such as processors, produce the majority of heat for the board. These components may become overheated if not provided with adequate airflow. Because the air flow is specified without factoring the location of the heat generating sources on the blade, it is often necessary to over-design the air flow generating elements which results in consumption of more-than-the-budgeted for power to drive the fans leaving less power for payload related operations. Moreover, power consumption is likely to migrate higher with increasing processor speed available to a server-blade based application. The limits of performance of the components in an ATCA blade may very well depend on its ability to dissipate the heat generated.
A circuit board may include an Intelligent Platform Management Controller (IPMC) that complies or is compatible with the Intelligent Platform Management Interface (IPMI) Standard detailed in “Intelligent Platform Management Interface Specification Second Generation,” Document Revision 1.0, dated Feb. 12, 2004, published by Intel, Hewlett-Packard, NEC, and Dell. The ATCA shelf may also include a shelf manager (hereinafter designated “ShMC”) to perform manageability functions for the shelf. Conventional shelf managers are typically implemented in software or firmware, or both. The shelf manager is typically capable of performing one or more manageability functions with respect to one or more elements populated on the shelf such as, for example, ATCA circuit cards, power modules, cooling units, field replaceable units and other resources and/or functionality of the shelf that are shared between all or a subset of the ATCA circuit boards. Typically the shelf manager communicates with one or more IPMCs comprised in one or more circuit boards. The shelf manager may reside on an ATCA circuit board that it manages or located in other circuit boards and/or other components populated on the shelf, or may be populated on shelf outside of the ATCA circuit boards or even lie external to the shelf.
The PICMG® Advanced Mezzanine Card (AdvancedMC or AMC) base specification, PIGMG AMC.0, Revision 1.0, published Jan. 3, 2005 (hereinafter referred to as the AMC.0 specification, the entire contents of which are incorporated herein by reference.) adds versatility to the modularity provided by the ATCA specification. The AMC specification defines the base-level mechanical, management, power, thermal, interconnect (including I/O) and system management requirements for hot-swappable, field-replaceable, add-on mezzanine cards (or modules) which may be hosted by an ATCA or a proprietary carrier board. Each AMC Module is received into an AMC Connector, seated parallel to the host carrier card and configured for high-speed, packet-based serial communications between the AMC card and the carrier board. There are six different form factors defined in the AMC specification which include two AMC module widths (W): the single width module (73.5 mm) and a double width module (148.5 mm); three heights (H) or thicknesses: compact (13.8 mm), mid-sized (18.96 mm) and full-sized (28.95 mm); and a single depth (D) (181.5 mm). The height (H) is measured in a direction normal to the major plane of the AMC card. The width (W) and height (H) dimensions lie along mutually perpendicular directions in a plane that is normal to the direction along which the depth (D) is measured. When the AMC module is mounted vertically, the width dimension is aligned vertically and the height or thickness dimension is aligned horizontally. The reverse is the case when the AMC module is mounted horizontally. Additionally, the AMC specification refers to three types of carrier board configurations—conventional, cutaway and hybrid.
The availability of AMC cards having a wide variety of form factors allows the cards to accommodate a rich mix of circuit elements and circuit topologies to support many different application architectures that can address the needs of diverse segments of the computer and telecommunications marketplace. The AMC architecture supports a number of transfer protocols with varying band widths as described in the subsidiary PICMG standard AMC3.0 for example. AMC cards extend the functionality of the ATCA carrier boards and permit multiple vendors to build technology solutions for transmission and switching equipment and allow these technology solutions to be used in multiple applications and in multiple vendor product lines. The ATCA standardization approach in general improves product reliability (allowing for industry standard hot swappable hardware and software, including power supplies and fans) and drives down prices—due in large part to greater economies of scale in manufacturing and less time spent on details standardized by ATCA (e.g., power, cooling, mechanical spacing and connectors issues).
Clearly, industry driven standards, such as the ATCA based standardization discussed above, allow performance-enhancing features and lower-cost assembly of systems. For example, in the case of the ATCA based standardization, the standard makes available standardized backplane, management functionality, power entry modules, cooling units and other modular components that are pluggable into a cost-reduced standardized-platform/chasses, prescribed by the standard, to deliver a user-configured solution in a logical package with lower component count and substantially simplified cabling and connectivity. However, technology implementations based on the ATCA specification represent “big iron” solutions that are suited to telephone company central offices with high density needs: i.e., switching systems and transmission cross connects. ATCA standardized chasses are too massive for remote/enterprise applications. Likewise, ATCA blades feature a form factor that makes them unsuitable for use in non-ATCA prescribed chasses such as the rack-mount chasses or shelf designed to be placed in the 19-inch standardized rack referenced above. It would be advantageous if the flexibility of a standards based architecture were made available in an enterprise class server blade offering.
The present invention is directed to a standards-based server blade arrangement wherein individual circuit boards may be compliant with a first industry driven or other standard and housed within an enclosure configured such that one aspect of the enclosure provides each circuit board with a scalable, mechanical, electrical and environmental functionality required for that circuit board to comply with the first industry driven or other standard and a second aspect of the enclosure allows the enclosure to comply with a second industry driven or other standard.
In one of the several possible embodiments, the present invention provides a reconfigurable enclosure that complies with mechanical and electrical features that allow each circuit board, which may conform to a first industry standard, to be removably received within a separate portion of the enclosure and placed into interconnected electrical relationship with the other circuit boards such that the collection of interconnected circuit boards and the enclosure comprise a server blade arrangement compliant with a second industry standard with each separate portion of the enclosure providing the electrical, mechanical and environmental functionality required by the particular circuit board housed within it to be compliant with the particular industry standard associated with that circuit board.
In another embodiment, the present invention provides an enclosure substantially shaped and dimensioned into a server blade form factor capable of being removably received within a standardized 19″ rack or shelf. One aspect of this embodiment includes an enclosure that is generally rectangular shaped with one pair of opposed sides configured with respective a front opening and a rear opening. An intermediate plane is disposed within the enclosure and transverse to the opposed sides on which the front and rear openings are located. The first and second openings are configured for receiving there-through respective first and second circuit cards and placing the cards in sliding engagement on guideways in the enclosure for opposed movement towards and removable coupling with attachment points on the intermediate plane. In another aspect of this embodiment, the intermediate plane provides the infrastructure to electrically and mechanically couple the first and second circuit cards to each other and couple the server blade to other server blades within the rack.
In another embodiment of the present invention, there is provided an enclosure comprising a front portion, an intermediate portion and a rear portion wherein the front and rear portions are adapted to receive and house respective first and second circuit cards therein. The intermediate portion is adapted to releasably mate with said front and rear portions so that the front and rear portions extend in opposed directions from the intermediate portion. One feature of this embodiment provides for the intermediate portion to remain detachably attached to the rack while each of the front and rear portions is removably coupled to opposed portions of the intermediate portion such that the front portion, the intermediate portion and the rear portion cooperatively comprise an enclosure substantially shaped and dimensioned into a server blade form factor that is capable of being removably received within a standardized rack or shelf such as, for instance, a 19″ rack or shelf. Another aspect of this embodiment is that each of the front and rear portions may be hot-swapped without substantially interfering with the operation of the server blade.
In accordance with another embodiment of the present invention, the intermediate portion is configured as a passive connector module that provides mechanical connection points for releasably anchoring a first and opposed second carrier boards and electrical conduction paths to electrically couple the first and second carrier boards without materially changing the signals in any manner.
An object of embodiments of the present invention is to provide an enclosure that conforms to a standards compliant server blade form factor that is constructed to allow access to the internal cavity housing the circuitry and mechanical components during operation and without the need to interrupt operation of the server blade.
The invention may be more completely understood in consideration of the following detailed description of various embodiments of the invention in connection with the accompanying drawings, in which:
Reference will now be made to the figures with reference to which the various features of the present invention will be described in detail. The drawings are generally not to scale and the visual perception of the dimensions of the various elements from the various drawings and figures is not intended to limit the invention in any way. In general, the same reference numeral is used to refer to the same element illustrated in separate drawings and/or separate views. The following description provides numerous specific details of the present inventions are set forth in terms of descriptions of exemplary embodiments framed in the context of the ATCA standard. However, it will be understood that this approach does not limit the use of the principles and teachings disclosed herein to ATCA equipment. One skilled in the relevant art will recognize that the invention can be practiced without one or more of the specific details, or with other methods, components, materials, etc. In other instances, well-known structures, materials, or operations are not shown or described in detail to avoid obscuring aspects of the invention.
The following embodiments are merely illustrative of one possible type of modular electronic device. In general, the principles and teachings are applicable to various types of modular electronic equipment, including, but not limited to, telecommunications equipment, data communication equipment and computer equipment. Also, it is understood that the phraseology and terminology used herein are for the purpose of description and should not be regarded as limiting. The use herein of “including” and “comprising” and variations thereof is meant to encompass the items listed thereafter and equivalents thereof, as well as additional items and equivalents thereof. Furthermore, the term “connected” is used herein to denote a direct physical and/or mechanical connection between elements. The terms “coupled,” “operably coupled”, or “operably connected” as used herein signify an indirect connection between elements.
Interior region 45 is partitioned into a right-cooling unit portion 50, a left-cooling unit portion 55, a frontal portion 60, a mid-portion 65 and a rear portion 70 best illustrated in
Referring now to
Reference is now made to
In the embodiment illustrated in
One embodiment of strut 155 will now be described with reference to
In one embodiment, strut 155 may also provide structural attachment points (not shown) to precisely locate and mount one or more of the backplanes 198, an extender card (alternatively midplane) 510, power modules 515 and shelf manager module 520 within the interior region 45 substantially proximate or within mid-portion 65 as will be described in the following sections. In alternate embodiments, guide channels may be formed on inner cover surfaces to slidingly receive and guide the strut 155 within the skeleton frame. Strut 155 provides structural support for the chassis 10 and the components installed within chassis 10. Strut 155 also provides a datum for the ATCA carrier cards and the backplanes mated to the ATCA carrier cards. Strut 155 thereby serves to position the backplanes 198 and the card guides 405 and 410 in precise geometrical relationship so that when the carrier cards 90(95) are installed within chassis 10, connectors 740, 745 and others present on the carrier card edge 710 are always positioned in aligned relationship with their counterparts on the backplane 198 to enable repeatable and error-free mating during the installation of ATCA carrier cards 90 (95) within chassis 10.
As illustrated in
In one embodiment of the present invention, cover 520 is placed over inner cover 150 of the skeleton frame 115 with edges 580, 585, 590 and 595 of cover 520 being in substantial parallel alignment with edges 180, 185, 190 and 195 of inner cover 150. Cover 520 is shaped and dimensioned such that cover-bottom surface 575 substantially conforms to a portion of the inner cover 150 such that at least one cover-aperture 515 is in substantial alignment with hole 220 on tab 200 so that a fastener can be inserted through each corresponding cover aperture 515 and hole 220 to releaseably fasten cover 520 to inner cover 150 of skeleton frame 115 as best illustrated in
As depicted in
Referring to
It will be appreciated that other ATCA-standard compliant architectures are also possible within the scope of the present invention. For example, the power module may be located within the mid-plane module 194 and connected to an external AC or DC power source via power rails that are ducted from the rear of the chassis 10 through interior region 45 and into the mid-plane region 194. Power from the power module is then provided to each of the power connectors on the individual ATCA carrier cards 90 (95) via the backplane and through the power connectors on the backplane that interconnect with the power connectors on the ATCA carrier cards when the cards are installed within chassis 10.
It must be emphasized that ATCA carrier cards are used for ease of description but not by way of limitation. The modular structure of the present invention allows for reconfiguring chassis 10 to be capable of accommodating circuit boards that comply with standards other than the ATCA specification with the mid-plane module providing the electrical and mechanical functionality to allow the circuit boards to interoperate within the chassis and within the rack in which the chassis is mounted. In alternate embodiments, the mid-plane module cooperates with one of the two circuit boards to provide appropriate electrical, mechanical and environmental functionality required for the proper operation of the other of the two circuit boards. In yet another embodiment of the present invention, one or more of the server blades in a rack may be specialized to deploy the electrical functionality—such as the power, fabric lanes, management and control to the remaining server blades in the rack.
Another feature of embodiments of the present invention is that a shelf manager may be advantageously housed within the extents of the mid-plane module 194 or optionally, the shelf manager implemented in firmware, may be located on one of the mezzanine cards.
In alternate embodiments, ATCA carrier card 95 is installed within chassis 10 but is configured to operate independently of ATCA carrier card 90 (i.e. the single board computer) while obtaining power and shelf manager services from the common resources housed within, the mid-plane module 194. In another embodiment, carrier card 95 may interconnect with carrier card 90 via the mid-place module 194. In this instance, carrier card 95 may function as a companion to carrier card 90 by supporting one or more storage devices and other peripherals via the use of mezzanine cards, for example, interconnecting with the carrier card 90 to provide RAID functionality to it. In another embodiment, several chassis 10 may be supported in a rack in stacked and interconnected relationship to each other. It will be appreciated that numerous other applications are possible using chassis 10 without deviating from the scope of the present invention. In one embodiment, the present invention provides a modular, scalable, highly reliable, readily reconfigurable, standards-based server blade system.
While the invention is amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and described in detail. It should be understood, however, that the intention is not to limit the invention to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims.
This application claims the benefit of U.S. Provisional Application No. 60/911,244 filed Apr. 11, 2007, which is hereby fully incorporated herein by reference
| Number | Date | Country | |
|---|---|---|---|
| 60911244 | Apr 2007 | US |
