U.S. patent application number 14/092254 was filed with the patent office on 2015-12-10 for method and apparatus for trusted branded email.
This patent application is currently assigned to JOSE J. PICAZO SEPARATE PROPERTY TRUST. The applicant listed for this patent is JOSE J. PICAZO SEPARATE PROPERTY TRUST. Invention is credited to Noel Ruane, Harish Seshadri.
Application Number | 20150358335 14/092254 |
Document ID | / |
Family ID | 34840451 |
Filed Date | 2015-12-10 |
United States Patent
Application |
20150358335 |
Kind Code |
A9 |
Seshadri; Harish ; et
al. |
December 10, 2015 |
Method and Apparatus for Trusted Branded Email
Abstract
A trusted branded email method and apparatus in one aspect
detects branded electronic messages and performs validation before
it is sent to a recipient. In another aspect, an electronic
messages is branded by embedding branding assets and validation
signatures. Algorithms that generate validation signatures are
dynamically selected to further strengthen the security aspects.
Branding assets are presented to a user using a distinct indicia
that represents to the user that the branding assets are
secure.
Inventors: |
Seshadri; Harish;
(Fullerton, CA) ; Ruane; Noel; (Palo Alto,
CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
JOSE J. PICAZO SEPARATE PROPERTY TRUST |
Livermore |
CA |
US |
|
|
Assignee: |
JOSE J. PICAZO SEPARATE PROPERTY
TRUST
Livermore
CA
|
Prior
Publication: |
|
Document Identifier |
Publication Date |
|
US 20140090044 A1 |
March 27, 2014 |
|
|
Family ID: |
34840451 |
Appl. No.: |
14/092254 |
Filed: |
November 27, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12234054 |
Sep 19, 2008 |
8621217 |
|
|
14092254 |
|
|
|
|
11035476 |
Jan 13, 2005 |
7457955 |
|
|
12234054 |
|
|
|
|
60536557 |
Jan 14, 2004 |
|
|
|
Current U.S.
Class: |
726/7 |
Current CPC
Class: |
H04L 9/3236 20130101;
H04L 2209/76 20130101; G06Q 10/107 20130101; H04L 2209/68 20130101;
H04L 2209/805 20130101; H04L 9/3247 20130101; H04L 2209/60
20130101; H04L 63/123 20130101; H04L 63/0428 20130101 |
International
Class: |
H04L 29/06 20060101
H04L029/06 |
Claims
1. An electronic messaging method comprising: an email detector
processor receiving an email message; the email detector processor
determining whether the email is a branded email message, wherein a
branded email message comprises embedded branding assets associated
with an email sender, and a generated algorithm index that
de-references the number of algorithms configured; if the email
message is a branded email message, the a configuration handshake
processor determining whether there is a configuration match
between a writer subsystem and a reader subsystem; if there is a
configuration match between the writer subsystem and the reader
subsystem, a policy binder processor obtaining a mapping between a
"from" domain name and a policy hash from the configuration; and if
the configuration match does not exist, the configuration handshake
processor passing the email message to an error handler process.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of U.S. Provisional
Patent Application No. 60/536,557 entitled Method and Apparatus for
Trusted Branded Email, filed on Jan. 14, 2004, which application is
incorporated herein by reference in its entirety.
TECHNICAL FIELD
[0002] The present application relates to processing branded
electronic mails in a secure manner.
BACKGROUND
[0003] The exchange of emails across organizations or from an
organization to an end user are widely accepted and promoted
business practices. B2B (Business to Business) email exchanges, and
the delivery of targeted email to end users, are practices that are
growing exponentially across business enterprises. Organizations
can promote their brand through these types of email campaigns. In
order to enforce the value of a brand and promote brand email
campaigns, email messages can include branding assets together with
the contents of the email. Branding assets are basically media
constructs like a company logo (image or movie), company sound
effects, or any other media that promotes a company's brand.
Moreover, branding assets can also be any additional artifacts used
for promoting business driven email campaigns. Branding assets or
references to the assets are typically embedded within the email
header. In order to reliably execute brand email campaigns, the
brand mail that includes branding assets and the message content
should be inviolable, for example, thereby enforcing the integrity
of the brand. By enforcing brand integrity, brand-specific
campaigns/scenarios can be executed adding to the value of an
organization's brand. However, if brand integrity is compromised,
there could be serious repercussions for an organization's brand,
for example, brand value deflation.
[0004] The brand mail system addresses these concerns by allowing
organizations to execute branding email campaigns reliably,
efficiently and securely. But why should BrandMail be secure? Brand
mail should be secure because of the importance of "brand-value"
for an organization or an entity. "Brand-Value" is enforced by
preserving the integrity of branded emails. The brand email
integrity is preserved by making the email tamper-proof and
unforgeable. If tampering or forging occurs the value of an
organization's brand may deflate significantly.
[0005] Organizations exchanging brand mail(s) can be assured that
both parties are truly communicating between each other (no forging
of a message) and that the mail contents have not been tampered
with. The same applies to the scenario where a branded mail is sent
from an organization (for example, hp.com) to an end-user
registered within a web mail system (for example, Yahoo, Hotmail).
In order for the brand mail system to be a reliable trust mechanism
it should address main security concerns.
SUMMARY
[0006] Method and apparatus for trusted branded email is provided.
In one aspect, the method includes receiving a branded electronic
mail and retrieving an algorithm index from the branded electronic
mail. An algorithm is dynamically selected using the algorithm
index, and the algorithm is used to generate a validation
signature. A tamper proof digest from the branded electronic mail
is retrieved and compared with the generated validation signature.
If the tamper proof digest matches the validation signature, the
branded electronic mail is transmitted to a recipient.
In another aspect, the method includes presenting the branded
electronic mail with a visual distinction to the recipient. The
visual distinction may be a branding asset associated with a sender
of the branded electronic mail such as a logo, moving images,
movie, or sound effect that, for example, identifies the
sender.
[0007] Yet in another aspect, the system in one embodiment includes
a writer subsystem that composes the secure branded email including
a tamper proof digest and a reader subsystem that validates the
tamper proof digest to ensure that the branded email has not been
tampered with.
[0008] Further features as well as the structure and operation of
various embodiments are described in detail below with reference to
the accompanying drawings. In the drawings, like reference numbers
indicate identical or functionally similar elements.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 is an architectural diagram illustrating a brand mail
system of the present application in one embodiment.
[0010] FIG. 2 is a topology diagram illustrating a secure brand
mail system of the present application in one embodiment.
[0011] FIG. 3 is a high-level diagram illustrating an
organization-to-organization use case of the present application in
one embodiment.
[0012] FIG. 4 is a high-level diagram illustrating an organization
to web mail end user use case of the present application in one
embodiment.
[0013] FIG. 5 is a flow chart illustrating a brand mail application
startup of the present application in one embodiment.
[0014] FIGS. 6A, 6B, and 6C are flow charts illustrating a
BrandMailReader of the present application in one embodiment.
[0015] FIGS. 7A, 7B, and 7C are flow charts illustrating a
BrandMailWriter of the present application in one embodiment.
DETAILED DESCRIPTION
[0016] The method and apparatus of the present application in one
embodiment includes a brand mail system that addresses the security
concerns such as forging, tempering, and cryptography in electronic
communications. A classic example of forgery is when a black hat
impersonates some organization or entity, and sends deceptive email
message(s) to an end user.
[0017] Briefly, a black hat describes a hacker who performs hacking
operations such as breaking into a computer system(s) or network(s)
with malicious intent. Hackers perform their operations ("crimes")
clandestinely. Unlike a white hat hacker, the black hat hacker
takes advantage of the break-in, perhaps destroying files or
stealing data for some future purpose. The black hat hacker may
also make the exploit known to other hackers and/or the public
without notifying the victim. This gives others the opportunity to
exploit the vulnerability before the organization is able to secure
it. They also forge data and tamper with existing data with the aim
of disrupting business processes. A white hat describes a hacker
who performs hacking for legitimate reasons. Typically white hats
are very qualified and contribute their knowledge to secure
existing systems.
[0018] By instilling a trust mechanism, parties exchanging emails
will be more confident about the source of the message(s). The
brand mail system addresses this problem, for example, by detecting
the forgery/tampering of an email message.
[0019] A classic "tampering-with" scenario is when a black hat
decides to intercept an email message and tamper with the contents
of the email. The black hat then could modify any part of a
message. This kind of tampering is known as a "man in the middle"
(MITM) attack. The tamper proof mechanism of the present
application in one embodiment addresses this problem, for example,
by validating the email message signature via a TPD thus enforcing
the data integrity of the message. TPD, in one embodiment, is a
base-64 encoded tamper proof digest, which is a brand validation
signature. In one embodiment, the tamper-proof mechanism (TPD
generation/validation) provides a trust mechanism for brand
mail.
[0020] There are many existing industry strength mechanisms that
perform email cryptographic operations such as SSL/TLS, S/MIME, and
PGP--(PKI solutions). The PKI solutions listed above can also be
used to secure brand mail messages. However, the present
application provides an optional high-performance lightweight
symmetric key cryptographic system based on dynamic algorithmic
selection based on a generated algorithm index. It is left to the
organizations communicating, for example, those exchanging brand
mails to configure their brand mail system for cryptographic
communication.
[0021] FIG. 1 is an architectural diagram illustrating an overview
of a brand mail system of the present application in one
embodiment. In one embodiment, a brand mail system of the present
application acts as a lightweight, robust, extensible, scalable,
and efficient trust mechanism. In one embodiment, a brand mail
system of the present application sits between conventional
communication systems such as an SMTP sender 110 and an SMTP
receiver 112. A brand mail system in one embodiment may include a
BrandMailWriter 106 and a BrandMailReader 108, subsystems
functioning to allow for the delivery of branded emails. In one
embodiment, the subsystems that perform the reading and associated
functions, and writing and associated functions are referred to as
BrandMailReader 108 and BrandMailWriter 106. However, the actual
names of the subsystems are not critical, rather any other
software, firmware or hardware systems or subsystems may be used to
perform the similar functions of the reader and the writer. A
BrandMailReader 108 in one embodiment is a subsystem generally
responsible for receiving an email message and detecting whether an
email message is branded or not. It may be a configurable email
proxy that receives either a regular email message branded message
114. If it is a regular email message, then the role of the
BrandMailReader is to redirect or forward the message 118 to a
designated SMTP receiver email server 112 within the respective
organization. If it is a branded message 114, the BrandMailReader
passes the branded mail through validation processors utilizing
branded mail configuration 104.
[0022] During the validation processing, errors may be generated
and handled. If the validation is successful, the message is
forwarded to a designated SMTP receiver email server 112 while
preserving the integrity of the branded mail.
[0023] A BrandMailWriter 106 in one embodiment is a subsystem that
is responsible for sending a branded email message 114. Utilizing
brand mail configuration 102, the BrandMailWriter 106, for example,
embeds references to branding assets or the assets themselves
within the branded email. The BrandMailWriter 106 also signs the
message with a TPD to ensure the integrity of the branded email
message. Optionally, it can also encrypt the email message body
with the generated cryptographic key.
[0024] The BrandMailWriter 106 in one embodiment is a configurable
email proxy that fabricates a branded email message from a regular
email message 116 (i.e. the emails are originally sent by email
users 120 or applications 122). The branded message is sent to the
BrandMailReader 108, then to the appropriate targets such as email
clients 124, web mail applications 126, and browsers 128.
[0025] The BrandMailWriter 106 and the BrandMailReader 104, in one
embodiment, utilize a brand mail cryptographic system for
optionally encrypting and decrypting email messages. A brand mail
cryptographic system in one embodiment is an optional,
configurable, lightweight symmetric key cryptographic system. In
one embodiment, encryption of a branded mail message happens within
the BrandMailWriter; decryption of the message happens within the
BrandMailReader. The BrandMailWriter uses the generated
cryptographic key to encrypt the mime body parts of the message.
The algorithm selected is based on the algorithm index generation
specific to the e-mail message. After the BrandMailWriter 106
composes an encrypted message, it ultimately sends the message to
the destination BrandMailReader 104. The BrandMailReader 104
detects the brand mail cryptographic tag, which indicates that the
message is encrypted and needs to be decrypted. Upon detection, the
BrandMailReader 104 decrypts the message content using the
algorithm index to select the decryption algorithm using the
generated cryptographic key. The BrandMailReader 104 then proceeds
to TPD validation. In one embodiment, if a cryptographic tag is not
detected, the BrandMailReader 104 proceeds directly to TPD
validation.
[0026] FIG. 2 is a topology diagram illustrating a secure brand
mail system of the present application in one embodiment. Emails
212 sent from organizations 202 to end users 206 typically pass
through MSPs 204. End users 206 receiving the emails 214 typically
view the messages using a browser 216 or email clients 218. The
brand mail system of the present application can reside between the
organization's email sending system 202 and end user's email system
and protect the messages from potentially dangerous black hats
hackers 208 by validating the messages.
[0027] FIG. 3 is a high-level diagram illustrating an
organization-to-organization use case of the present application in
one embodiment. In this example, a sender from organization X sends
a branded mail to a receiver from organization Y. For instance, a
sender using known electronic mail systems 302 can write an email
message and send the email to an organization Y's receiver. The
brand mail system 304 of the present application in one embodiment,
for example, using a BrandMailWriter and BrandMailReader
subsystems, which will be explained in detail below, transforms the
regular email into branded mail and delivers it to the email
communication endpoint. At 306, organization Y's receiver views the
branded email that the organization X's sender sent. The present
application provides a visual distinction, for example, in the form
of a company brand logo (in the user's inbox) that indicates that
the e-mail is secure. This shows that by displaying a trademark
also conveys a trustmark (since the e-mail is secure). The
BrandMail system is capable of enacting any visual distinction
anywhere within the email so that it represents a trustmark. The
visual distinction provided by the present application may be any
indication that would convey to a user that the email is secure,
for example, that it passed a security check. Some examples of the
visual distinction may include a particular color indication
associated with the email in an inbox, a particular picture or
symbol, or even a predetermined sound, tactile or textual indicia
associated with the email. Indicia may be included in a particular
field of an inbox display or presentation. The particular field of
an inbox may be an existing field or a newly added field used for
the indicia. Thus, when a user sees or hears (or feels) the
particular indicia associated with the email, the user knows that
the email is secure.
[0028] In one embodiment, when using third party email clients, a
plug-in/COM add-in may be created to process the branding assets.
In one embodiment, the plug-in/COM add-in is the visual processing
component of the BrandMail system. It can securely render the
branding assets (i.e. displaying the logo) of the branded e-mail.
The plug-in/COM add-in can read the headers of new mail and based
on the fundamental brand mail headers, for example, certain
predetermined fields of the brand mail headers, can determine that
it is a brand mail. The plug-in/COM then can read the standard
visual branding assets such as bm_logo and bm_sound effect, for
example, contained in the brand mail header.
[0029] Generally, branding assets refer to any asset that enforces
a brand such as a logo (for example, a company logo) or sound
effects (audio) that represent a company. The logo can either be an
image or a movie (for example, a QuickTime movie such as Intel
Inside) or both. Branding assets can also include criteria that
promote branding business processes such as additional assets for
logging, metrics analysis, filtering, rating, rules, and anti-spam
tactics. Brand assets are usually embedded within the email header,
but can optionally be embedded as a mime body part. The brand asset
format usually is preceded by a brand mail asset prefix (for
example, bm_). A brand asset header entry can contain resources,
and/or references to resources. The resources can either be text
resources or network reference resources. Text resources can be any
tag containing text values. For example text resources can be
descriptions, relative paths, digests as text, meta-tags, formatted
rules, etc.
[0030] An example of a text resource is a brand asset that provides
a description of a company's mission statement. The header entry is
shown below: [0031] bm_mission_statement=To become a global leader
in the industry for email CRM development On the other hand network
reference resources provide references to network resources.
Examples of such are hyperlinks (URLs/URIs), RDF, JNDI, Web
services, etc. An example of a network resource is a brand asset
that represents a company's logo. A logo is a major depiction of a
brand, thus logos are displayed to represent the company and its
brand. The header entry is shown below: [0032]
bm_logo=www.hp.com/media/images/logo.gif [0033] or [0034]
bm_logo_movie=www.hp.com/media/movies/logo.mov
[0035] In one embodiment, header assets such as bm_logo (image) can
be registered as a standard visual brand mail asset and every brand
mail will be checked for the existence of a secure logo header
entry. A method and system of the present application in one
embodiment, first checks to see if it can find a bm_logo header
entry. If the entry exists then it uses it as the logo asset. If
the logo asset is not found then the system knows that it is not
dealing with a branded mail.
[0036] After reading the branding assets, the plug-in/COM can pull
in the associated media, for example, by making the URL (universal,
resource locator) requests to the paths provided for the brand
mail. This case implies that the email client (for example, Eudora,
Outlook) user is online and connected to the Internet. In one
embodiment, each time the user connects to the Internet, the media
is updated and cached locally for the given domain. This way, even
if a end-user using an email client such as Outlook is not online,
an end user still can get to experience the cached media such as
logos and sound effects.
[0037] FIG. 4 is a high-level diagram illustrating an organization
to web mail end user use case of the present application in one
embodiment. As shown at 402, organization X's sender writes an
email and sends the email to web mail user receiver. At 404, the
brand mail system of the present application in one embodiment
transforms the regular mail into branded mail and delivers it to
the email communication endpoint, for instance, using
BrandMailWriter and reader subsystems. At 406, the web mail user
receiver views the branded email that organization X's sender
sent.
[0038] In one embodiment, a web mail application 406, for instance
a third party web mail application, which typically makes use of
dynamically generated web pages, gets all mail messages for the
particular folder. If there are any brand mail messages, the web
mail application's protocol can be changed from http to https.
Optionally, the web mail application can always run https. In one
embodiment, for each brand mail message, the web mail application
may run any domain specific rules, for example, implemented as Java
script that executes any other code fragments like jsp (JavaServer
Pages), asp (Active Server Pages) page fragments, which may be
embedded as branding assets in the e-mail header, on the branding
assets if applicable, and display a logo and process any other
media such as playing sound effects, if applicable. In one
embodiment, brand assets such as the logo are pulled in to the view
by making requests to the bm_logo or bm_logo_movie URL. FIG. 4
illustrates an inserted logo ("DELL") 408 with the received brand
mail. FIG. 4's logo 408 is an example of a visual distinction
provided in the present application. A display of the logo 406 in
the inbox as shown, for example, manifests to a user that this
branded email is a secure message, so that the user can safely open
the message. Companies are assured that their logo has not been
tampered or otherwise hacked.
[0039] FIG. 5 is a flow chart illustrating a brand mail application
startup of the present application in one embodiment. At 502, when
a brand mail application subsystem such as a BrandMailWriter or a
BrandMailReader starts up, a brand mail configuration XML file is
validated. The XML file validation generally checks to see if all
the necessary configuration nodes exist, and that they contain
valid values. It also checks, for example, the integrity of XML via
DOM (Document Object Model and XML schema validation. At 504, if
the configuration is invalid, the application logs the
configuration errors and exits gracefully. If the configuration is
valid, at 506, the configuration data is loaded, for instance, into
efficient memory data structures for access during the run time
mail processing. The brand mail configuration may include system
wide configuration constructs as well as domain specific
configuration constructs.
[0040] The valid domain specific configurations are hashed and
stored for the given domain, in one embodiment, each domain
configuration node in the XML is converted into a
DomainConfigurationObject. A mapping is created in memory that
associates this domain configuration with its corresponding hash.
This mapping is also stored, for instance, within an efficient
memory data structure for run-time mail processing access. Aspects
such as the ordering of hash classes, cryptographic key generation
classes and cryptographic classes are preserved by the hashing of
the domain specific configuration.
[0041] A domain specific configuration, for example, contains a
mapping of domain names to attributes that are domain specific. The
domain specific configuration in one embodiment contains a base-64
encoded policy hash, defines brand assets specific to the domain,
hashing, cryptographic key generation, and cryptographic algorithms
(for example, all indexed with the natural order of the definition
(the order that the cryptographic algorithm definitions appear in
the configuration file), a flag for using optional constructs such
as cryptography, and log directory configurations for valid/invalid
TPD emails.
[0042] Below is an example of the XML domain configuration that is
representative of the domain specific configuration object
model:
TABLE-US-00001 <domain name="intel.com"> <policy
value="c3d6a2ffb6e4c6a3d744d2e8b24fa1da"/> <brand-assets>
<brand-asset name="bm_logo"
value="http://www.intel.com/images/intel_logo.gif "/>
<brand-asset name="bm_sound_effect"
value="http://www.intel.com/sounds/intel_sound.wa v"/>
</brand-assets> <hash-classes> <hash-class
name="com.brandmail.algorithms.hash.MD5Hasher"/> <hash-class
name="com.brandmail.algorithms.hash.SHA1Hasher"/>
</hash-classes> <use-cryptography value="true"/>
<cryptographic-key-generation-classes> <class
name="com.brandmail.algorithms.cryptography.keyge
neration.ModulusKeyGenerator"/> <class
name="com.brandmail.algorithms.cryptography.keyge
neration.DivisionKeyGenerator"/>
</cryptographic-key-generation-classes>
<cryptographic-classes> <cryptographic-class
name="com.brandmail.algorithms.cryptography.Tripl eDES"/>
<cryptographic-class
name="com.brandmail.algorithms.cryptography.BlowF ish"/>
</cryptographic-classes> <tpd-valid-directory
name="/intel.com/TPD/valid/"/> <tpd-invalid-directory
name="/intel.com/TPD/valid/"/> </domain>
[0043] Other domain configurations are possible. For example,
additional definitions or nodes may be added to the domain
configuration or the existing ones modified.
[0044] Configuration file contains a configuration for the
subsystem of interest. It defines a subsystem type that should be
configured as for example, BrandMailReader or BrandMailWriter. It
also encapsulates all the domain specific configurations. Thus, for
example, a BrandMailReader or BrandMailWriter subsystems access
their respective configuration files. Below illustrates an example
configuration:
TABLE-US-00002 <configuration> <subsystem
type="BrandMailWriter"/> <domain-specific-configuration>
<domain name="intel.com"> . . . </domain> . . .
</domain-specific-configuration> </configuration
[0045] Like the domain specific configuration, more configuration
nodes can be added or the existing ones modified.
[0046] In one embodiment, the brand mail system of the present
application ensures that the brand mail algorithm index
dereferences the right index for hashing, cryptographic key
generation, and/or cryptography between the BrandMailWriter and
reader systems. For example, in the case of cryptography, if the
algorithm index de-referenced different non-corresponding
algorithms (for the writer and reader respectively), the decryption
of the mail would be invalid due to the BrandMailReader/writer
configuration mismatch. That is, the writer would encrypt with a
specific algorithm class while the reader would decrypt using a
different algorithm class. Therefore, the data would be designated
as corrupted, and there may not be an easy way to discern a
configuration mismatch from a tampering/forging scenario.
[0047] In order to make the distinction between a configuration
mismatch from a tampering/forging scenario, the brand mail system
of the present application in one embodiment checks for the
equality of the configuration hashes of BrandMailReader and writer.
In one embodiment, the domain specific configuration hash is used
to validate the BrandMailWriter's configuration against the
BrandMailReader's configuration. In one embodiment, the
BrandMailWriter writes its configuration hash into the mail headers
and the BrandMailReader validates its configuration against
BrandMailWriter's value. If the values match then processing
continues otherwise errors are logged and processing stops.
[0048] In one embodiment, the brand mail system of the present
application provides multiple domain specific algorithms, for
example, hash algorithms, key generation algorithms, cryptographic
algorithms, for instance, to increase the security of the system.
In one embodiment, every branded email, which may have a good
degree of uniqueness, generates an algorithm index (checksum),
which is used to de-reference the number of algorithms configured.
As the number of algorithms configured increases, the
unpredictability of what algorithm is chosen (by the algorithm
index) for the given branded email also increases. The degree of
unpredictability is attained by performing the following modulus
operation: [0049] [algorithm_index]%[# of registered algorithm
classes] This unpredictability makes it difficult for black hats to
break the brand mail system. To further better the security of the
system, it is recommended to tweak or add code to the existing
industry-strength algorithms in a manner that is reliable and
efficient. By doing this, black hats cannot count in having the
exact algorithm for breaking the system.
[0050] FIGS. 6A, 6B, and 6C are flow charts illustrating a
BrandMailReader of the present application in one embodiment. At
602, a message that is sent, for example, by the BrandMailWriter is
received by the BrandMailReader email proxy. At 604, a
BrandMailDetector Processor performs a detection that diagnoses the
following information: whether the mail message is branded message
or just simply a regular message. The detection in one embodiment
is performed by reading the header entries and seeing if the email
headers contain all of the fundamental brand mail headers. The
fundamental brand mail headers in one embodiment include the
following: [0051] bm_domain_config header entry [0052] bm_tpd
header entry [0053] bm_index header entry [0054] bm_written_time
header entry [0055] bm_logo header entry
[0056] In one embodiment, all of the fundamental brand mail header
entries must exist (both the key/value pairs of the header entries)
for the mail to be considered as a branded mail at 606. If all the
fundamental brand mail header entries exist then the
processing/validation of the brand mail continues. If none of the
fundamental brand mail headers exist then the reader determines
that it is dealing with a regular message. At 608, regular mail
messages are simply redirected to the adequate SMTP receiver
through the Transport Processor. At 609, if the send was
successful, the processing stops. Otherwise, appropriate error
handling and logging is performed at 611 and 613 for the
message.
[0057] There is a case where only some of the fundamental brand
mail headers are embedded. In one embodiment, if this is the case
then it is determined that the message has been tampered-with.
Consequently the processing for the message stops with appropriate
error handling at 610 and the message is logged at 612 as invalid
brand mail. Other methods to detect whether a message is a branded
mail can be utilized, including, for example, setting a flag,
setting a minimum number predetermined fields, etc.
[0058] If the message is a branded mail, at 614,
ConfigurationHandshake Processor makes sure that the domain
configuration of the writer matches the domain configuration of the
reader. This is done by checking the brand mail header
"bm_domain_config" field value against the domain specific
configuration hash of the reader.
[0059] The domain specific configuration hash for the reader is,
for example, generated and stored on application startup. For
example, the domain configuration hashes were created during server
startup of the reader and writer and the hashes are the same for
the email sender and email receiver to exchange information. If at
616 there is a configurations match then the processing continues
to 618, otherwise the processing branches to a BrandMailException
indicating misconfiguration, and performs error handling at
610.
[0060] At 618, a PolicyBinder Processor gets the mapping between
the "from" domain name and the base-64 encoded policy hash from the
brand mail configuration and subsequently binds the policy hash as
a mail processing attribute. In one embodiment, the policy hash
designates the contractual agreement between both parties
exchanging e-mails. This makes the communication secure between
both parties. In one embodiment, it is used as the core symmetric
key. In one embodiment, a default implementation makes use of a XML
file in the file system. Below is an example of the XML node that
is configured in the brandMailConfiguration.xml file:
TABLE-US-00003 <domain name="hp.com"> ... <policy
value="b8a3021de82f3e01db57b9aec4d31252"/> ... </domain>
<domain name="support.hp.com"> ... <policy
value="c3d6a2ffb6e4c6a3d744d2e8b24fa1da"/> ... </domain>
<domain name="dell.com"> ... <policy
value="82f3e01db57b9ab6e4c6a3d76a2ffb6e"/> ...
</domain>
[0061] The PolicyBinder processor at 618, in one embodiment, grabs
the policy hash using the "from" domain as the key. This is done,
for example, by invoking the method "String bindPolicy( . . . ),"
which returns the policy hash of interest or throws a
PolicyNotFoundException if a policy hash is not found. The policy
hash entry is then bound as a mail attribute. The binding in one
embodiment is a mapping record between the current mail processing
thread, the mail mime message, and a map of attributes.
[0062] If at 620, a policy hash is not found, the processing
branches to 610 for error handling and exits. If at 620, a policy
hash that maps to the domain is found, the processing continues to
CryptographicDetector at 622 shown in FIG. 68.
CryptographicDetector Processor, in one embodiment, in the context
of the BrandMailReader, employs a cryptographic detector that
invokes the method "boolean shouldApplyCryptographyForReader( . . .
)". This method checks to see if the "bm_cyph" header entry exists
and its corresponding value is set to true. If this is the case the
processing continues for cryptographic decryption, otherwise
cryptography does not apply.
[0063] Thus, at 624, if cryptography is used, for instance,
"bm_cyph" header entry with its corresponding value set to true was
found, the processing continues to 626. Otherwise, the processing
branches to TPDValidator Processor at 628.
[0064] At 626, CryptographicKeyGenerator Processor generates a
cryptographic key and binds it as a mail attribute. The
cryptographic key will then be used by the cryptographer processor
at 530. In one embodiment, the algorithm selection for a
cryptographic key is based on the algorithm index ("bm_index"
header entry). There is a collection of registered cryptographic
key generation algorithms as classes implementing the
CryptographicKeyGenerator interface. In one embodiment, the classes
are configured for each domain. Below is an example of the XML node
that is configured in the brandMailConfiguration.xml file:
TABLE-US-00004 <domain name="hp.com"> ...
<cryptographic-key-generation-classes> <class
name="com.brandmail.algorithms.cryptography.keyge
neration.DivisionKeyGenerator"/> <class
name="com.brandmail.algorithms.cryptography.keyge
neration.ModulusKeyGenerator"/>
</cryptographic-key-generation-classes> ...
</domain>
[0065] In one embodiment, the same order of the domain specific
cryptographic key generation classes is maintained within the
configuration of the BrandMailWriter. If this does not happen there
may be an algorithm mismatch between the writer and the reader. In
order to select the true algorithm index in the collection the
following modulus operation is performed: [0066] bm_index % [# of
registered cryptographic key generation classes] This gives the
true index to dereference the appropriate cryptographic key
generation class within the collection. The method "byte[ ]
generateCryptographicKey( . . . )" is invoked by the selected
cryptographic key generation class and the return value
representing the generated key is bound as a mail attribute for the
respective mail request. The mail attribute is called
"cryptographic_key". This key is subsequently used by the
Cryptographer processor to perform any cryptographic
operations.
[0067] At 630, a Cryptographer Processor decrypts the message body
parts. In one embodiment, it uses the "cryptographic_key" mail
attribute as the key to decrypt the mime body parts. The algorithm
selection in one embodiment is based on the algorithm index
("bm_index" header entry). There is a collection of registered
cryptographic algorithms as classes implementing the Cryptographer
interface. The classes should be configured for each domain. Below
is an example of the XML node that is configured in the
brandMailConfiguration.xml file:
TABLE-US-00005 <domain name="hp.com"> ...
<cryptographic-classes> <class
name="com.brandmail.algorithms.cryptography.BlowF ish"/>
<class name="com.brandmail.algorithms.cryptography.Tripl
eDES"/> </cryptographic-classes> ... </domain>
[0068] In one embodiment, the domain specific cryptographic classes
are maintained within the configuration of the BrandMailWriter.
Otherwise, there may be an algorithm mismatch between the reader
and the writer. In order to select the true algorithm index in the
collection the following modulus operation is performed in one
embodiment: [0069] bm_index %[# of registered cryptographic
classes] This gives the true index to dereference the appropriate
cryptographic class within the collection. Once the cryptographic
class is selected, the "byte [ ] decrypt( . . . )" method is called
for all the mime body parts. In case of any problems a
BrandMailException is thrown and errors are processed via the
ErrorHandler processor. At 632, if the decryption was successful,
the processing continues to TPDValidator Processor at 628.
Otherwise, the processing branches to ErrorHandler Processor at 634
and appropriate logging at 638.
[0070] The TPDValidator Processor at 628 validates the TPD for the
message. In one embodiment, to validate the TPD, this processor
makes use of the following constructs in one embodiment: [0071]
Algorithm index (The "bm_index" header entry which is embedded in
the email headers) [0072] The collection of registered TPD
algorithms as classes implementing the Hasher interface. The
classes should be configured for each domain. Below is an example
of the XML node that is configured in the
brandMailConfiguration.xml file:
TABLE-US-00006 [0072] <domain name="hp.com"> ...
<hash-classes> <class
name="com.brandmail.algorithms.hash.SHA1Hasher"/> <class
name="com.brandmail.algorithms.hash.MD5Hasher"/>
</hash-classes> ... </domain>
[0073] in one embodiment, the same order of domain specific hash
classes are maintained within the configuration of the
BrandMailWriter. If this does not happen there may be an algorithm
mismatch between the writer and the reader. In order to select the
true algorithm index in the collection the following modulus
operation is performed in one embodiment: bm_index %[# of
registered hash classes]
This provides the true index to dereference the appropriate hash
class within the collection. Because the algorithm index is not the
same for all message, the algorithm thus selected rotates among the
algorithms listed in the domain configuration file.
[0074] A concatenation of the algorithm index (which has a good
degree of uniqueness) and the data of each mime body part is
performed. The result of the concatenation is then hashed based
using the adequate algorithm de-referenced by the algorithm index.
The generated value is compared with the "bm_tpd" value found in
the email header. At 636, if the values match the validation is
successful then the processing continues to a TPDLogger processor
at 642 in FIG. 6C, otherwise the validation errors are logged at
638 and the system halts further mail processing for this message
at 640.
[0075] At 642, the TPDLogger Processor logs the TPD in a file
system. In one embodiment, the name of the file is the base-64
encoded TPD. The contents of file is the complete email message.
This is used to keep a log of all the brand mails processed for
audit purposes. The configuration of the directory path is provided
in the brand mail xml configuration. Below is an example of the XML
node that is configured in the brandMailConfiguration.xml file:
TABLE-US-00007 <domain name="hp.com"> ...
<tpd-valid-directory name="/hp.com/TPD/valid/"/>
<tpd-invalid-directory name="/hp.com/TPD/invalid/"/> ...
</domain>
[0076] In one embodiment, to discern whether the brand mail system
is dealing with a valid TED, a boolean attribute called
"tpd-validity" is checked. The BrandMailReader in one embodiment
implements the TPDValidator processor to bind this boolean email
attribute with a "tpd-validity" value. In one embodiment, this
Boolean attribute is bound to the mail object. If the TPD validity
flag is set to true, then the brand mail system creates a TPD file
entry with the complete contents of the email within the
"tpd-valid-directory." If the TPD validity is false, then
TPDValidation failed for the email, so a TPD file entry is created
with the complete contents of the email within the
"tpd-invalid-directory." If the "tpd-validity" flag is missing then
the processing stops because there are errors before TPD
generation.
[0077] At 622, if the TPD is valid, the processing continues to a
Transport processor at 646. If the TPD validity is not found, the
processing stops at 654. At 646, the Transport processor sends the
branded mail to the SMTP receiver server. At 648, if the send was
successful, processing stops at 654. However, if the send failed,
the processing branches to ErrorHandler at 650. At 650, the
ErrorHandler Processor generically handles the errors of
processors. In the cases of failure, the processors add error
attributes to the errors map. The errors that were populated within
the map are logged by the error handler at 652.
[0078] FIGS. 7A, 7B, and 7C are flow charts illustrating a
BrandMailWriter of the present application in one embodiment. In
one embodiment, when the BrandMailWriter application starts up, the
brand mail configuration XML file is validated. If the
configuration is invalid the application logs the configuration
errors and exits gracefully. If the configuration is valid, the
configuration data is loaded into efficient memory data structures
for access during the run time mail processing.
[0079] In one embodiment, the brand mail configuration includes
system wide configuration constructs as well as domain specific
configuration constructs. The valid domain specific configurations
are hashed and stored for a given domain. This mapping is also
stored within an efficient memory data structure for run-time mail
processing access. Aspects such as the ordering of hash classes,
cryptographic key generation classes and cryptographic classes are
preserved by the hashing of the domain specific configuration since
the brand mail algorithm index de-references the index for hashing,
cryptographic key generation, and cryptography between the
BrandMailWriter and reader systems.
[0080] For example, in the case of cryptography, if the algorithm
index dereferences different non-corresponding algorithms for the
writer and reader respectively, the decryption of the mail would be
invalid due to the BrandMailReader/writer configuration mismatch.
For instance, the writer would encryption with a specific algorithm
in the writer while the reader would decrypt with a different
algorithm. Therefore, the data would be deemed corrupted and
furthermore there would not be an easy way to discern a
configuration mismatch from a tampering/forging scenario. In order
to make the distinction between a configuration mismatch from a
tampering/forging scenario, the brand mail system of the present
application checks for the equality of the configuration hashes of
BrandMailReader and writer.
[0081] In one embodiment, a domain specific configuration hash is
used to validate the BrandMailWriter's configuration against the
BrandMailReader's configuration. The BrandMailWriter writes its
configuration hash into the mail headers and the BrandMailReader
validates its configuration against BrandMailWriter's value. If the
values match then processing continues otherwise errors are logged
and processing stops.
[0082] At 702, the BrandMailWriter in one embodiment detects or
receives an incoming regular message. As an example, a regular
message is sent by an email client for instance via webmail,
outlook, eudora, etc. and received by the BrandMailWriter email
proxy, where for example, the host name and port # is configured by
the email client.
[0083] At 704, a RemoteAddressInNetwork Processor, for instance,
checks whether the remote address (the address of the email client)
in the network. For example, in one embodiment, permissible
domains, addresses, subnets are configured within the
BrandMailWriter.
[0084] If the email client is in the network, the processing
continues to 706. Otherwise, processing branches to ErrorHandler
Processor at 708. At 706, a DomainConfigurationHashWriter
Processor, for instance, writes a header entry corresponding to the
domain specific configuration hash. An example header entry is
listed below:
[0085] bm_domain_config=<Base-64 domain specific configuration
hash>
[0086] In one embodiment, domain specific base-64 configuration
hashes are generated for all the domain configurations on
application startup. The "from" email address domain, for example,
appearing in a conventional electronic message fields, is used to
lookup the corresponding configuration hash and subsequently the
value is stored as a header entry. If the domain is not found at
708, a DomainNotFoundException is thrown at 710.
[0087] At 712, a PolicyBinder Processor for instance, gets the
mapping between the "from" domain name and the base-64 encoded
policy hash from the brand mail configuration and subsequently
binds the policy hash as a mail processing attribute. In one
embodiment, the default implementation makes use of a XML file in
the file system. Below is an example of the XML node that is
configured in the brandMailConfiguration.xml file:
TABLE-US-00008 <domain name="hp.com"> ... <policy
value="b8a3021de82f3e01db57b9aec4d31252"/> ... </domain>
<domain name="support.hp.com"> ... <policy
value="c3d6a2ffb6e4c6a3d744d2e8b24fa1da"/> ... </domain>
<domain name="dell.com"> ... <policy
value="82f3e01db57b9ab6e4c6a3d76a2ffb6e"/> ...
</domain>
[0088] The PolicyBinder processor at 712, in one embodiment, grabs
the policy hash using the "from" domain as the key. This is done,
for example, by invoking the method "String bindPolicy( . . . ),"
which returns the policy hash of interest or throws a
PolicyNotFoundException if a policy hash is not found. The policy
hash entry is then bound as a mail attribute. The binding in one
embodiment is a mapping record between the current mail processing
thread, the mail mime message, and a map of attributes.
[0089] At 714, if policy exists, the processing continues to
AssetAdder Processor at 716. Otherwise, the processing branches to
ErrorHandler Processor at 710 and appropriate logging at 711. At
716, the AssetAdder Processor adds the brand mail assets to the
email headers by calling the "void addAssets( . . . )" method. In
one embodiment, the brand mail assets are found within the brand
mail configuration. The key is the "from" domain and values are
sub-nodes representing the assets. Below is an example of the XML
node that is configured in the brandMailConfiguration.xml file:
TABLE-US-00009 <domain name="hp.com"> ...
<brand-assets> <asset name="bm_logo"
value="http://www.hp.com/images/hp_logo.gif"/> <asset
name="hp_filter" value="B2B,marketing campaign,upsell"/>
<asset name="hp_rating" value="classified"/>
</brand-assets> ... </domain> <domain
name="intel.com"> ... <brand-assets> <asset
name="bm_logo" value="http://www.intel.com/images/intel_logo.gif
"/> <asset name="bm_sound_effect"
value="http://www.intel.com/sounds/intel_sound.wa v"/>
</brand-assets> ... </domain>
[0090] In one embodiment, after adding the brand mail assets an
additional header tag indicating the time the assets are written to
the headers is inserted. The inserted tag's name is
"bm_written_time" and is a string representation of the time in
milliseconds in one embodiment. Referring to FIG. 7B, at 718 if
there are problems adding the brand mail assets then a
BrandMailException is thrown and the processing branches to
ErrorHandler Processor at 720 and to TPDLogger Processor at 721 for
appropriate error handling and logging.
[0091] It is noted that any brand asset can be defined for the
given domain as long as the brand asset name is unique for the
given domain. In one embodiment, a logo brand asset (for example,
"bm_logo") is a standard visual brand asset that is added in the
configuration.
[0092] At 722, an AlgorithmIndexGenerator Processor generates an
algorithm index, for example, by concatenating the policy hash
(retrieved from the mail attributes map), and all the other header
entries. The result is then run through a checksum algorithm such
as the CRC-32. What is returned is a long value that is then
converted to a string. An example of a method used to accomplish
this is a "Long generateAlgorithmIndex( . . . )". The Long checksum
is converted to a string and stored within the email headers as the
following header entry:
[0093] bm_index=<long checksum value as string>
The algorithm index has a degree of uniqueness based on the policy
hash, and the header entries. The algorithm index is used to select
the right algorithm to generate the TPD. At 724, if there are
problems generating the algorithm index, then a BrandMailException
is thrown and the processing branches to 720 and to TPDLogger
Processor at 721 for appropriate error handling and logging. If on
the other hand, an algorithm index was generated successfully, the
processing continues to TPDGenerator Processor at 726.
[0094] TPDGenerator Processor in one embodiment is responsible for
generating a TPD for the message. To generate a TPD for the
message, this processor in one aspect makes use of the following
constructs: [0095] Algorithm index (The "bm_index" header entry
which is embedded in the email headers) [0096] The collection of
registered TPD algorithms as classes implementing the Hasher
interface. The classes are configured for specific domains. Below
is an example of the XML node that is configured in the
brandMailConfiguration.xml file:
TABLE-US-00010 [0096] <domain name="hp.com"> ...
<hash-classes> <class
name="com.brandmail.algorithms.hash.SHA1Hasher"/> <class
name="com.brandmail.algorithms.hash.MD5Hasher"/>
</hash-classes> ... </domain>
[0097] In one embodiment, the same order of domain specific hash
classes are maintained within the configuration of the
BrandMailReader. If this does not happen there may be an algorithm
mismatch between the writer and the reader. In order to select the
true algorithm index in the collection the following modulus
operation is performed in one embodiment: bm_index % [# of
registered hash classes]. Thus, the algorithm is dynamically
selected using the algorithm index. Because the algorithm index is
not the same for all message, the algorithm thus selected rotates
among the algorithms listed in the domain configuration file. This
provides the true index to dereference the appropriate hash class
within the collection.
[0098] A concatenation of the algorithm index (which has a good
degree of uniqueness) and the data of each mime body part is
performed. The result of the concatenation is then hashed based
using the adequate algorithm de-referenced by the algorithm index.
The hash called TPD is base-64 encoded and stored as a header entry
called: bm_tpd=<TPD> Additionally a boolean mail attribute
called "tpd-validity" is bound and set to true. This will be used
by the TPDLogger.
[0099] At 728, if a TPD was generated successfully, the processing
continues to CryptographicDetector Processor at 730. If the TPD
generation failed, the processing jumps to an ErrorHandler
Processor at 720 and to TPDLogger Processor at 721 for appropriate
error handling and logging. At 730, the CryptographicDetector
Processor employs a cryptographic detector to invoke a method
"boolean shouldApplyCryptographyForWriter( . . . )". This method
first checks to see if the brand mail configuration includes a tag
for cryptography for the given domain. Below is an example of the
XML node that is configured in the brandMailConfiguration.xml
file:
TABLE-US-00011 <domain name="hp.com"> ...
<use-cryptography value="true"/> ... </domain>
[0100] If the "use-cryptography value" node is not provided or set
to false, then cryptography is not used. Referring to FIG. 7C, if
at 732, cryptography used, the processing branches to a
CryptographicKeyGenerator Processor at 734. Otherwise, the
processing continues to a TPDLogger Processor at 733.
[0101] At 734, the CryptographicKeyGenerator Processor generates a
cryptographic key and binds it as a mail attribute. The
cryptographic key will then be used by the cryptographer processor
at 740. In one embodiment, the algorithm selection is based on the
algorithm index ("bm_index" header entry). There is a collection of
registered cryptographic key generation algorithms as classes
implementing the CryptographicKeyGenerator interface. The classes
are configured for each domain. Below is an example of the XML node
that is configured in the brandMailConfiguration.xml file:
TABLE-US-00012 <domain name="hp.com"> ...
<cryptographic-key-generation-classes> <class
name="com.brandmail.algorithms-cryptography.keyge
neration.DivisionKeyGenerator"/> <class
name="com.brandmail.algorithms.cryptography.keyge
neration.ModulusKeyGenerator"/>
</cryptographic-key-generation-classes> ...
</domain>
[0102] In one embodiment, the same order of the domain specific
cryptographic key generation classes are maintained within the
configuration of the BrandMailReader. If this does not happen there
may be an algorithm mismatch between the writer and the reader. In
order to select the true algorithm index in the collection the
following modulus operation is performed: bm_index [# of registered
cryptographic key generation classes]
This provides a true index to dereference the appropriate
cryptographic key generation class within the collection.
[0103] For example, a method "byte[ ]generateCryptographicKey( . .
. )" is invoked by the selected cryptographic key generation class
and the return value representing the generated key is bound as a
mail attribute for the respective mail request. The mail attribute
is called "cryptographic_key". This key is subsequently used by the
Cryptographer processor to perform any cryptographic
operations.
[0104] At 740, using the generated key, the Cryptographer Processor
performs the encryption of the message. The Cryptographer
Processor, for example, is responsible to encrypt the message body
parts. It uses the "cryptographic_key" mail attribute as the key to
encrypt the mime body parts. The algorithm selection is based on
the algorithm index ("bm_index" header entry). There is a
collection of registered cryptographic algorithms as classes
implementing the Cryptographer interface. The classes are
configured for each domain. Below is an example of the XML node
that is configured in the brandMailConfiquration.xml file:
TABLE-US-00013 <domain name="hp.com"> ...
<cryptographic-classes> <class
name="com.brandmail.algorithms.cryptography.BlowF ish"/>
<class name="com.brandmail.algorithms.cryptography.Tripl
eDES"/> </cryptographic-classes> ... </domain>
[0105] In one embodiment, the same order of the domain specific
cryptographic classes are maintained within the configuration of
the BrandMailReader. If this does not happen there may be an
algorithm mismatch between the writer and the reader. In order to
select the true algorithm index in the collection the following
modulus operation is performed: bm_index % [# of registered
cryptographic classes] The algorithm is thus, dynamically selected,
for example, to further enhance the security mechanism.
[0106] This gives the true index to dereference the appropriate
cryptographic class within the collection. Once the cryptographic
class is selected, the "byte [ ]encrypt( . . . )" method is called
for all the mime body parts and finally a header entry is added
indicating that cryptography was performed. The header entry is
listed below: bm_cyph=true
[0107] If cryptography is not used then it is not necessary to add
the above tag. At 742, if the encryption procedure was failed, a
BrandMailException is thrown and the processing branches to 736 for
appropriate error handling. If the encryption procedure was
successful, the processing continues to 733.
[0108] At 733, a TPDLogger Processor logs the TPD in the file
system. The name of the file is the base-64 encoded TPD. The
contents of fire is the complete email message. This is used to
keep a log of all the brand mails processed for audit purposes. The
configuration of the directory path is provided: in the brand mail
xml configuration. Below is an example of the XML node that is
configured in the brandMailConfiguration.xml file:
TABLE-US-00014 <domain name="hp.com"> ...
<tpd-valid-directory name="/hp.com/TPD/valid/"/> ...
</domain>
[0109] To determine whether a tpd is valid, a boolean attribute
called "tpd-validity" is checked against. In one embodiment of the
BrandMailWriter flow, the TPDGenerator processor will always set
"tpd-validity=true", since the writer is not validating the TPD. If
the TPD validity flag is set to true, then a TPD file entry is
created with the complete contents of the email. The
tpd-valid-directory configuration is used to write the TPD log
contents to the configured location. If the "tpd-validity" flag is
missing then the processing stops, because there are errors before
TPD generation.
[0110] With a valid TPD, the processing continues to Transport
Processor at 744. Transport Processor dispatches the created
branded mail to the BrandMailWriter's SMTP sender service
component. At 746, if the send is successful, the processing stops
at 748. Otherwise, the processing branches to an ErrorHandler
Processor at 736.
[0111] The ErrorHandler Processor at 736 generically handles the
errors of processors. In the cases of failure the processors add
error attributes to the errors map. The errors that were populated
within the map are logged by the error handler, for instance by a
TPDLogger Processor at 738.
[0112] The system and method of the present disclosure may be
implemented and run on a general-purpose computer. The embodiments
described above are illustrative examples and it should not be
construed that the present invention is limited to these particular
embodiments. Thus, various changes and modifications may be
effected by one skilled in the art without departing from the
spirit or scope of the invention as defined in the appended
claims.
* * * * *
References