WO2003009174A1

WO2003009174A1 - Improvements in or relating to database access security

Info

Publication number: WO2003009174A1
Application number: PCT/IL2002/000575
Authority: WO
Inventors: Moshe Raab
Original assignee: Log-On Software Ltd.
Priority date: 2001-07-16
Filing date: 2002-07-16
Publication date: 2003-01-30
Also published as: US20060059154A1; EP1407385A1; US7904454B2; EP1407385A4

Abstract

Apparatus for protection of database objects from unwanted access, particularly from external connections via a firewall (20). The apparatus comprises a data packet parsing unit (54) for parsing a data packet to find database operation commands in the packet, and an enforcement unit (56) for applying enforcement rules to the data packet, thereby to protect respective database objects. The apparatus may form an additional layer (50, 52) of protection in conjunction with a firewall (20) to protect internal data.

Description

Improvements in or relating to Database Access Security

Field of the Invention

The present invention relates to database access security and more

particularly but not exclusively to a method or apparatus for securing database

material against unauthorized access from an external network whilst at the

same time permitting legitimate access from the external network. Examples

of database systems of interest are MS-SQL Server, Oracle, DB2 and UDB.

Background of the Invention

Networks typically require users to log on, that is give usernames and

passwords to gain access. Based on the username the network may decide on a

level of access. Local Area Networks are typically connected to the outside

world and often have a firewall to protect the perimiter of the LAN from

external intrusions. The firewall may permit users to log in from outside the

LAN. Such users are hereinafter referred to as external users, by contrast with

internal users, whose connections are from within the perimeter of the LAN.

The firewall may apply restrictions to such external users, or may not allow

them at all, and the firewall may also monitor and control e-mail, web-

interaction and the like. Typically, in the case of the external user, the firewall

may check the username at the periphery and decide what access level to

provide. The firewall may specify particular machines on the LAN to which the externally connecting user is permitted access, or it may provide general

access, but once access to a machine is granted, the access is unrestricted. The

firewall does not monitor, validate or in any other way consider actual usage,

except in the case of widely known protocols, as discussed below.

Today, an organization's information systems are often prime targets for

vandalism and theft, and organizations are especially concerned about the

activities performed by external users. Loss or corruption of a database can be

catastrophic for an organization, and information in the database can be of use

to rivals. A database that includes customers' credit card numbers, or staff

bank account numbers, can be of great interest to thieves. Currently, external

users are able to log in, and the resulting access connection is assigned use

rights associated with the logged in user. On the one hand, the organization

wishes to allow staff to be able to work from home or have access to the LAN

whilst traveling etc. yet at the same time free external access gives rise to

threats of the kind described above. It is a goal of firewall technology to

provide maximal levels of security for minimal levels of inconvenience, and

one way of achieving this goal is to target access restrictions as closely as

possible to the kind of activity which constitutes the threat, leaving non-threat

bearing activity to continue unimpeded.

To reduce possible threats, organizations are therefore moving to

implement application specific security that can monitor the actual activity of

external users and block activity that is suspicious, thereby to protect against

intrusion from outside the LAN that is application specific. Software packages that provide data use surveillance already exist on the market. However, the

existing packages are restricted to well known and standardized protocols such

as HTTP, SMTP and FTP. These packages are able to identify the use of the

particular protocol and use their knowledge of the protocol to monitor and

control the content. Databases however, do not use any single recognized

protocol. Indeed most database packages are not accompanied by any

published protocol and most database versions, even from the same source,

include variations in the protocol used. There is thus no application specific

security package that is currently aimed at databases, and it is not currently

possible to carry out content based monitoring and control of database queries

or results. .

Summary of the Invention

An object of the present invention is to provide a data security layer that

monitors and/or controls database access statements or commands, for example

the actual SQL statements that constitute much database activity for users and

application servers. As mentioned above, typical database systems include MS- SQL Server, Oracle, DB2 and UDB.

According to a first aspect of the present invention there is thus provided

apparatus for protection of database objects from unwanted access comprising:

a data packet inspection unit for inspecting passing data packets to find

and carry out an analysis of database operation text within said packet, and an enforcement unit, associated with said data packet inspection unit

for applying enforcement rules to said data packet, based at least partly on said

analysis, thereby to protect respective database objects.

Preferably, said data inspection unit comprises a packet analysis unit to

look for structure associated with database operation text, and a parsing unit,

associated with said search unit to parse said database operation text into

underlying statements comprising at least database operation commands and

database objects.

Preferably, said data packet inspection unit is configurable in association

with a firewall between an external network and an internal LAN containing

said database objects, thereby to guard a respective database.

Preferably, said inspection unit is configurable to be positioned between

two LANs.

Preferably, said data packet inspection unit is further operable to find

information regarding sources of respective data packets and regarding data

objects of respective commands.

Preferably, said enforcement rules each comprise at least one condition,

said condition being the presence in the database operation text of a

predetermined database operation command.

Preferably, said enforcement rules each comprise at least one condition,

said condition being the presence in the database operation text of any one of a

group of database operation commands belonging to a preformulated group. Preferably, said wherein said enforcement rules each comprise at least

one condition, said condition being the specification within the database

operation text of a predetermined database object.

Preferably, said enforcement rules each comprise at least one condition,

said condition being the specification, within the database operation text of any

one of a group of database objects belonging to a preformulated group.

The apparatus may further comprise a management module for setting

said enforcement rules.

Preferably, said management module is further operable to define

groups of database commands, said groups being usable to define said rules.

Preferably, said management module is accessible via a graphical user

interface.

Preferably, said management module is operable to set an access policy

comprising a plurality of rules, each rule specifying a group of commands, a

group of users and a group of data objects, said access policy being transferable

to said rule application unit.

Preferably, at least some of said enforcement rules define logging

activities and said enforcement unit further comprises logging functionality

therefore. Preferably, said inspecting comprises obtaining a signature from said

database operation text and wherein at least some of said rules relate at least in

part to said signatures.

According to a second preferred embodiment of the present invention

there is provided a method for protection of database objects from unwanted

access comprising:

parsing a data packet to find database operation commands in said

packet, and

applying enforcement rules to said data packet, said rules specifying

conditions based at least partially on said database operation commands,

thereby to protect respective database objects.

The method may further comprise finding information regarding sources

of respective data packets and regarding data objects of respective commands.

Preferably, said finding information regarding sources of respective data

packets is carried out per user connection, the method further comprising

associating said sources with data packets of said user connection.

The method may further comprise applying said enforcement rules

selectively to said data packets, depending on information found in respective

data packets by said parsing units.

Preferably, said database operation commands are arranged in groups

and wherein said rules relate to said groups. Preferably, said command sources are arranged in groups and said rules

relate to said groups.

Preferably, said database objects are arranged in groups and wherein

said rules relate to said groups.

The method may further comprise arranging said database commands into

groups and using said groups to define said rules.

The method may further comprise forming an access policy by

arranging together a plurality of rules for use.

Brief Description of the Drawings

For a better understanding of the invention and to show how the same

may be carried into effect, reference will now be made, purely by way of

example, to the accompanying drawings.

With specific reference now to the drawings in detail, it is stressed that

the particulars shown are by way of example and for purposes of illustrative

discussion of the preferred embodiments of the present invention only, and are

presented in the cause of providing what is believed to be the most useful and

readily understood description of the principles and conceptual aspects of the

invention. In this regard, no attempt is made to show structural details of the

invention in more detail than is necessary for a fundamental understanding of

the invention, the description taken with the drawings making apparent to those skilled in the art how the several forms of the invention may be embodied in

practice. In the accompanying drawings:

Fig. 1 is a generalized schematic diagram illustrating a general concept

of a first preferred embodiment of the present invention,

Fig. 2 is a simplified schematic diagram showing a second preferred

embodiment of the present invention, in which an SQL filter is located inwards

of a DMZ, the DMZ having a database query application,

Figs. 3A, 3B and 3C are simplified schematic diagrams of three

versions of the embodiment of Fig. 2, in which the elements of the embodiment

of Fig. 2 are mapped onto a series of servers,

Fig. 4 is a simplified schematic diagram showing in greater detail the

SQL filter of Figs. 1-3,

Fig. 5 is a generalized screen shot showing a dialog window for setting

up a rule for the filter of Fig. 4,

Fig. 6 is a generalized schematic diagram showing network

configuration for the SQL filter of Fig. 4 and illustrating a management console

for programming the filter,

Fig. 7 is a generalized screen shot showing a dialog window for

establishing a security policy for the filter of Fig. 4 by insertion of rules,

Fig. 8 is a simplified flow chart showing the process of creating a

policy for use in the filter of Fig. 4. Description of the Preferred Embodiments

The present embodiments provide a system for monitoring and or

controlling database command statements as they pass in to a LAN from an

external connection. The commands used may be monitored by the

embodiments as well as the source of the commands and the data objects to

which the commands apply. Rules and policies define legitimate and

illegitimate activity by using the various commands, sources and objects, so

that for example different grades of data can be protected from different types

of uses by different users or categories of user. In the present embodiments,

protection is not provided by access permissions assigned following log-in

operations, but rather by the actual data as found in the individual data packet,

in particular by the database access or operational commands, especially SQL

commands. Thus, non-database activity is not affected by the embodiments

and furthermore, database restrictions can be made as specific as possible so as

to target threats with as little effect as possible on legitimate database activity.

The embodiments described herein are for use together with a firewall,

although additional embodiments that may be used on their own to provide

database protection are contemplated.

Before explaining at least one embodiment of the invention in detail, it

is to be understood that the invention is not limited in its application to the details of construction and the arrangement of the components set forth in the

following description or illustrated in the drawings. The invention is applicable

to other embodiments or of being practiced or carried out in various ways.

Also, it is to be understood that the phraseology and terminology employed

herein is for the purpose of description and should not be regarded as limiting.

Reference is now made to Fig. 1, which is a simplified schematic

diagram showing a generalized embodiment of the present invention. In Fig. 1,

a database 10 is located on a LAN 12. The LAN 12 is protected by a firewall

14, through which external users 16 are able to connect in order to use the

database. External users who successfully negotiate the firewall, are typically

able to make use of the database, and the firewall is typically able to grant or

refuse access to the computer on which the database is located. The firewall

itself is not able to provide database protection that is more sophisticated than

providing access or no access. Thus, according to the present embodiments,

there is provided an additional layer of protection 18, referred to hereinbelow

as an SQL filter, which operates regardless of permission levels granted by the

firewall 14, and which checks passing data packets for SQL or other database

commands. If it finds database commands then it uses its own rules, again

independent of the firewall, to decide whether the data packet should be

allowed to pass. However, in current embodiments the system administrator

has the responsibility of ensuring that all incoming packets pass through the

filter. The firewall 14 is shown as a two-part firewall, having a first firewall

part 20, and a second firewall part 22 with a demilitarized zone, (DMZ) 24 in

between. The demilitarized zone typically includes a web server 26 and an

application server 28. In one configuration, if it is desired to permit outside

users 16 to access database 10 then an application, that is to say a database

client, for using the database, is placed in the DMZ 24. The external user

connects to the database client and uses the database client to generate SQL

code to access the database. Thus, data packets directed inwardly from the

firewall, and being intended to interrogate or manipulate the database, may be

expected to carry SQL or like database operating commands. The SQL filter

18 preferably snoops data packets passing between the firewall and the LAN,

identifies and analyzes any SQL material in the packets and applies rules to

decide whether the packets may pass or not. The rules are preferably

independent of those applied by the firewall.

Reference is now made to Fig. 2, which is a representation in block

diagram form of an embodiment operative in accordance with the present

invention, and which is similar to the generalized embodiment of Fig. 1, with

some minor variations. Parts that are the same as in Fig. 1 are given the same

reference numerals and are not referred to again except as necessary for an

understanding of the present embodiments. In Fig. 2 the SQL filter 18 is

shown to be part of the inner firewall 20. The server 26 is shown to support an

SQL application and the SQL guard checks data packets passing the inner

firewall and either authorizes them, in which cases they proceed, or it rejects them, in which case the packets are dropped. The DMZ 24 carries a database

client application 26, as before so that an external user may connect to the

client in the DMZ and issue database commands. The issued database

commands are forwarded to the inner firewall 20.

. At the inner firewall 20, which constitutes the entry point to the user

organization's LAN 12, the SQL-filter obtains the traffic to the database, and

only such traffic. If other protocols attempt to pass through SQL-filter they are

preferably rejected. The SQL filter assumes that all data traffic towards the

database from the outside is directed through the filter by other network

devices, and it is the responsibility of the network administrator to make sure

that there is no alternative unsecured path to the database.The filter analyzes

the data stream and finds user names of the externally connecting users, and

any SQL statements within the data packets. The proxy holds a security policy,

as will be explained in more detail below. The policy comprises a list of

security rules for accepting or rejecting the packets. The rules preferably

determine authorization levels on a packet by packet basis, and may take into

account such information as the user name, the SQL commands in the packet

(e.g. select, update) and the database objects (e.g. a table name, view name) to

which the SQL is applied.

The SQL-filter 20 of the embodiment of Fig. 2 is a proxy to the inner part

of the firewall and receives therefrom notifications on new TCP connections to

the database servers. The filter 20 preferably obtains all of the data stream of

the database server protocol for screening. The SQL-filter may authorize or forbid any database request in accordance with the rules as currently set, as

discussed above.

A currently available prototype is provided as an add-on to Firewall- 1 by

Check Point. Using the OPSEC (Open Platform for Security) by Check Point,

the firewall is induced to tunnel the relevant data stream to the SQL-filter,

which is then able to perform its function of filtering the database commands.

The prototype is suitable for any kind of database system, but in particular the

MS-SQL Server, Oracle Server, DB2 and UDB)

Reference is now made to Figs. 3 A, 3B and 3C, which are simplified

server block diagrams showing the arrangement of Fig. 2 mapped to a series of

servers. In the configuration of Fig. 3 A, a first server 40 holds the DMZ and

the database application, a second server 42 holds the inner firewall and the

SQL guard, and a third server 44 holds the database that the application in the

first server 40 is intended to query. The configuration of Fig. 3B is the same as

that of Fig. 3 A except that the application within the DMZ routes SQL-bearing

data packets to the SQL filter, rather than this being the responsibility of the

firewall. In Fig. 3C, the firewall once again has the responsibility of routing

SQL-bearing data packets to the filter, however the filter itself is located in the

safe zone within the firewall.

The arrangement of Fig. 3 shows the SQL-filter as a stand-alone proxy,

whether on the firewall server 42 or within the LAN. In accordance with any

of the above configurations, the database, on the database server 44, is only

able to receive data that has been approved by the filter. SQL-filters according to various of the present embodiments can

analyze data streams on any relational database technology, e.g. MS-SQL

Server, DB2, and UDB by IBM. Filters according to the preferred

embodiments require only the following, that a User name or like identification

can be identified, and that requests are sent as SQL statements. Other

embodiments may relate to other systems of database commands and the SQL

requirement may be modified accordingly.

Reference is now made to Fig. 4, which is a simplified diagram showing in

greater detail an SQL filter 18 suitable for the embodiments of Figs. 1-3. As

discussed above, the filter monitors packets passing through the inner firewall

20, and comprises four internal units, a session management unit 50, a packet

analyzer 52, an SQL parser 54 and an SQL validator 56. The internal units are

now considered in turn. First of all the session management unit 50 performs

overall management of individual connecting sessions. The session manager is

responsible for holding data for each TCP connection or session, including IP

addresses, user name, and internal control data, hereinafter referred to as

session data. For each incoming packet in the data stream, the session manager

associates the appropriate session data as a context for any SQL statement

contained within the packet.

The packet analyzer 52 analyzes packet content to find SQL text within

the packet. SQL text is passed to the SQL parser to identify the commands and

the data objects to which it relates, and then all of the information obtained is

passed on to the SQL validator which determines whether the commands are authorized for the identified data object for the user identified in the

corresponding session data. For convenience in the formulation of rules, the

rules used by the validator do not usually relate to individual commands or

users or data objects but rather to groups of commands, groups of users and

groups of objects, as will be explained in greater detail below.

The packet analyzer 52 is responsible for identifying SQL messages in the

data stream. The TCP connection is a stream-wise connection. However SQL

statements, which constitute the database requests, have specific structure,

attributes and length, which may be identified.

The packet analyzer is responsible for isolating the database requests in the

data stream, and then analyzing as follows:

• On initialization - the packet analyzer obtains the source and target

addresses, the protocol version, and the connected user. This data is placed

in the session manager.

• In the course of the session, the packet analyzer identifies messages that

contain SQL text, including underlying statements or requests, and isolates

the SQL text. The SQL text is passed to the SQL parser 54 for parsing and

authorization checking.

In a further preferred embodiment the packet analyzer also checks other

types of messages, for example the results of the query. For example it is

possible to define the filter to prevent an external user from downloading more than a certain number of items from the database. In this embodiment, the

packet analyzer checks outgoing packets as well as incoming packets.

Database vendors tend to adapt database protocol structures for their own

purposes, as will be discussed in more detail below, and the analyzer is

preferably designed or programmed to accord with the protocol specification as

used by the vendor of the database server in question. In a company that uses

multiple databases from multiple vendors the analyzer is preferably is able to

determine, typically from the session manager, which database is being

interrogated, and to select the correct protocol for use in parsing.

The SQL parser 54 receives data from the packet analyzer 52 as described

above, and then is responsible for extracting the commands and objects that

appear in the SQL text within the packets. The parser 54 preferably operates by

breaking down the text of the SQL messages into underlying statements.

Individual underlying statements may trigger more than one authorization

action from the filter. For example a single statement may contain multiple

objects, each needing separate authorization, or the statement may by its

structure require several authorization actions. For example a single statement

may gather information from several sources to be shown in a single view.

Each source may be required to be authorized separately.

In the prototype referred to hereinabove, the SQL parser is based on LEX

and YACC technology, which are available both commercially and as

freeware. SQL keywords are treated as tokens and are predefined to the LEX

environment. The SQL statement format is likewise defined to the LEX environment. Generally SQL definition uses an ANSI standard, however

individual vendors tend to make significant modifications, mainly in the DDL

(Data Definition Language) part of the language. Individual implementations

preferably follow the database vendor language reference to ensure that the

filter gives full coverage of all the SQL dialect of the database being protected,

as discussed above in respect of the analyzer.

While parsing the language the parser collects commands and

corresponding objects from the data stream. Each command and object

combination is checked immediately, together with the corresponding session

data to determine whether the user is authorized or not for the requested

database action defined by the command and object combination. If the user is

not authorized, the entire SQL text portion under investigation is preferably

rejected, and as a further sanction the session with the user may be ended.

Authorizations are checked via the SQL Validator 56, as described below.

The SQL validator 56 receives validation requests from the parser and

carries out validation based on the information supplied by the parser.

Validation requests are of the form: is user "SCOTT" allowed to carry out

command "SELECT" from data object "EMP". On each identification of an

SQL command and object in the parsing process, a validation request is issued

to the validator 56.

The validation is carried out using a set of rules arranged in a data structure

which is optimized to enable efficient finding of the highest priority rule that matches the authorization request. Typically this means that the data structure

is arranged as a hierarchy.

On initialization, the SQL validator 56 loads a list of rules that belong to a

currently set policy, and constructs an optimized structure therefrom. At

predetermined intervals, or optionally upon occurrence of a predefined event,

the validator 56 may check whether a new security policy has been set.

Upon receipt of the validation request, the data to be validated is

matched against the optimized structure. The structure allows for rapid

selection of the highest priority rule that fits the request. Following application

of the appropriate rule, the packet is either allowed to pass or is discarded, and

the event may be logged. All acts of discarding may be logged, or certain

packets regarded as marginally suspicious may be logged but allowed through,

as required by the user. The filter may conveniently make use of the log 58 of

the firewall, or of a special log mechanism or of a private log.

The rules that are used are preferably grouped together into sets, and

the sets of rules are the policies referred to herein. Policies can be exchanged

to reflect changing security priorities, or for any other purpose.

Reference is now made to Fig. 5, which is a simplified diagram

showing a dialog window 60 for setting up a rule. The dialog window has a

plurality of fields for data entry. A policy name field 62 allows the rule to be

associated with a policy, namely a group of rules. It will be appreciated that a

single rule can be featured in a plurality of policies. A second field 64 allows the rule to be given a name, so that it can be referred to. A third field 66 allows

a group of database operation commands to be specified. The database

operation commands are typically SQL commands, e.g. Select, Create Table.

The groups are preferably set in advance by the user for convenience in setting

the rules. Thus, it may be useful to group database editing commands into one

group, and database management commands into another group.

A fourth field 68 specifies an object or group of objects of the database

command. The objects are the database objects referred to by the database

commands, for example an SQL object name, e.g. table name, index name, and

even stored procedure, and the group of objects is a group of such items. A

fifth field 70 allows a group of users to be specified, and a sixth field 72

specifies the action that is to be taken if a data packet fits the rule. Typical

actions, as discussed above, are accepting the packet, rejecting the packet, and

an action known as "drop" which comprises both rejecting the packet and

terminating the session, typically a TCP/IP session. A check box 74 allows the

user to define whether the action should be logged.

Two exemplary rules are as follows:

Allow, User=SCOTT, Command=SELECT, Object=EMP

Reject, User=OTHER, Command=CREATE TABLE, Object=N/A

In the above rules individual users and commands are specified.

However, it is not expedient to have to specify a separate rule for each

combination of users and commands and objects. To ease the definition of rules - a user interface allows the administrator to define groups of users,

groups of commands and groups of objects. A rule may typically be of the

format 'allow', <user group>, <command group>, <object group>. The

implication is a Cartesian combinations of all members of all the groups in the

rule. It will be appreciated that the scopes of rules may be extended by the use

of additional parameters , such as defining the time of the day when

authorization is allowed.

The rule applies the action defined therein to any data packet that

attempts to pass the filter. The filter is indeed similar to content filters used in

current firewall technology, the difference being that the filter of the present

embodiments uniquely looks for and recognizes the structure of SQL text.

Preferably, the filter is configured so that the default action for

unrecognized packets is rejection, so that only those packets specifically

recognized by a rule are allowed through. The filter thus behaves as if it has a

default rule at the base of the rules hierarchy applying to all groups of

commands, all users and all data objects and whose action is to reject. Only

those packets recognized by a rule having an "accept" action higher up in the

hierarchy are therefore going to be accepted. It is possible to change the

default behavior by adding a rule at the base of the hierarchy applying to all

groups and whose action is to reject and log, or reject and deny service.

It is noted that the use of default behavior means that when a new

database version is installed, any new commands of the new version are not

initially recognized and are therefore rejected. The filter is preferably updated with the new commands as soon as possible so that the commands can be

accepted, however, until that happens, the default behavior ensures that

modifications of the database do not lead to a hole in the security that the filter

provides.

In a further preferred embodiment of the present invention, the analysis

of the data packet involves obtaining a signature of the database text found

therein. The rules used by the enforcement module then relate to the signature.

Reference is now made to Fig. 6, which is a simplified block diagram

showing how a filter system may be configured for a network. Parts that are the

same as those in previous figures are given the same reference numerals and

are not referred to again except as necessary for an understanding of the present

embodiment. Preferably, the filter system comprises two component parts,

both of which are configured to work together with the network firewall. A

first part is a policy editor 80, which located within a firewall management

console 82, alongside a firewall management server application 84, and which

is placed behind the firewall 20 in the secure region of the user's internal

network. The second part is the filter 18 itself, which is located together with

the firewall 20.

It is noted that in Fig. 6, the accessing client application is located

externally, by contrast with the situation shown in Fig. 1, wherein the client

application was located in the DMZ. The location of the client application,

which produces the database commands, is not critical to the present embodiments provided that the data packets produced by the client program

pass the filter 18.

The policy editor 80 allows for rules to be formulated and edited and built

into policies which may then be put into operation within the filter 18. The

policy editor also allows for the arrangement of users, data objects and

commands into groups, and for subsequent editing of the groups. The policy

editor also allows for editing and arranging of policies themselves, and

reference is now made to Fig. 7, which is a simplified diagram showing a

policy dialog window. The policy dialog window shows a series of rules (rule

1, rule 2 etc) that have been chosen for inclusion within a policy. Each rule is

given a priority, and one of the reasons for giving a priority is that there is no

requirement that the rules are mutually exclusive. Thus, in the event of more

than one rule applying, the filter has to know which rule is to be applied. The

order of appearance of the rule in the policy rule defines its priority. The first

rule that is correctly matched with the requested authorization, is the one that

defines the authorization, and no subsequent matching is preferably carried out.

Reference is now made to Fig. 8, which is a simplified flow chart

showing the procedure for defining rules and policies. One or more database

servers that it is desired to protect are firstly given defining identities.

Secondly, users are defined, and thirdly groups are defined, and individual

users are assigned to the groups. Optionally external connecting locations can

be defined as trusted locations. The skilled person will appreciate that users are

most effectively grouped according to access level requirements and security issues raised. Fourthly, the database command groups are assigned and

individual commands are assigned to the groups. Fifthly, database objects are

defined. Sixthly object groups are defined and the individual objects are

assigned to groups. Finally individual rules are defined by specifying an object

group, a user group and a command group and assigning an action thereto. The

action can be modified by further defining whether it should be tracked or

logged. Policies may then be defined by selecting rules into a hierarchy, and

finally a most appropriate policy is selected for current use by the filter.

The embodiment described herein preferably provides a security layer,

additional to that of the firewall, for protecting databases by filtering and

monitoring SQL requests that are generated outside of the database server's

local area network. The preferred embodiments scan all TCP/IP packets that

are to be delivered to the database server, tracking each database command and

authorizing or rejecting each command, based on a table of one or more

policies, each comprising a set of rules, and which may be provided by the

user, or which may be supplied as defaults. The policies establish which users

can have access to the database, which commands they can use, and for which

data objects in the database. It is noted that only one policy may be active at

one time. Multiple policies are allowed within the system, but for different

SQL-filter locations or different policies for different days. The latter may be

desirable for example to permit remote maintainance on a specific day in the

month. The SQL-filter adds an additional layer of security to a database or

database system by checking SQL statements that cross the firewall from

external networks. Such statements are requied by the filter to conform to

authorization policies set by the system administrator to ensure that security

breaches, such as compromised database passwords, misconfigured database

security, database vulnerabilities etc. cause minimal damage. Using the present

embodiments, it is possible to give a user a certain level of access when

connecting within the organization, however to restrict his database access

when connecting from outside of the organization. Thus any given user may

connect from the outside and carry out non-database functions as he chooses.

However, as soon as he attempts to carry out database access functions then the

filter can be used to limit his activity independently of the firewall. It is thus

possible to assign users with full rights to a database when connecting

internally, and to assign to the same users full rights to connect from outside,

and additionally, in accordance with the present embodiments to specifically

restrict database access to any desired extent for any given user when

connecting from the outside. Prior art firewalls are able to define that certain

machines (identified by IP address) do not have access to computers that store

the database but do not in fact inspect for or in any way restrict actual database

usage.

Thus, the policies that can be set by the SQL-filter enable the system

administrator to differentiate between access permissions granted to users when

working locally and permissions they may be assigned when working externally. In addition, it is possible to use the information held in incoming

data packets to determine where the external connection comes from. Thus

certain external machines can be predefined as safe machines, or safe machines

for given users. Policies, which are set based on user's id, SQL commands,

and database objects, can thus also incorporate connecting locations.

The SQL-filter of the preferred embodiments thus preferably protects the

databases within the LAN from unauthorized external access and ensures that

unauthorized access attempts do not cross the firewall.

It is appreciated that certain features of the invention, which are, for

clarity, described in the context of separate embodiments, may also be provided

in combination in a single embodiment. Conversely, various features of the

invention which are, for brevity, described in the context of a single

embodiment, may also be provided separately or in any suitable

subcombination.

Unless otherwise defined, all technical and scientific terms used herein

have the same meanings as are commonly understood by one of ordinary skill

in the art to which this invention belongs. Although methods similar or

equivalent to those described herein can be used in the practice or testing of the

present invention, suitable methods are described herein.

All publications, patent applications, patents, and other references

mentioned herein are incorporated by reference in their entirety. In case of

conflict, the patent specification, including definitions, will prevail. In addition, the materials, methods, and examples are illustrative only and not

intended to be limiting.

It will be appreciated by persons skilled in the art that the present

invention is not limited to what has been particularly shown and described

hereinabove. Rather the scope of the present invention is defined by the

appended claims and includes both combinations and subcombinations of the

various features described hereinabove as well as variations and modifications

thereof which would occur to persons skilled in the art upon reading the

foregoing description.

Claims

1. Apparatus for protection of database objects from

unwanted access comprising:

a data packet inspection unit for inspecting passing data packets to find

and carry out an analysis of database operation text within said packet,

and an enforcement unit, associated with said data packet inspection unit

analysis, thereby to protect respective database objects.

2. Apparatus according to claim 1, wherein said data

inspection unit comprises a packet analysis unit to look for structure associated

with database operation text, and a parsing unit, associated with said search

unit to parse said database operation text into underlying statements comprising

at least database operation commands and database objects.

3. Apparatus according to claim 1 , wherein said data packet

inspection unit is configurable in association with a firewall between an

external network and an internal LAN containing said database objects, thereby

to guard a respective database.

4. Apparatus according to claim 1, wherein said inspection unit is

configurable to be positioned between two LANs.

5. Apparatus according to claim 1, wherein said data packet

inspection unit is further operable to find information regarding sources of

respective data packets and regarding data objects of respective commands.

6. Apparatus according to claim 5, wherein said enforcement

rules each comprise at least one condition, said condition being the presence in

the database operation text of a predetermined database operation command.

7. Apparatus according to claim 5, wherein said enforcement

the database operation text of any one of a group of database operation

commands belonging to a preformulated group.

8. Apparatus according to claim 5, wherein said wherein said

enforcement rules each comprise at least one condition, said condition being

the specification within the database operation text of a predetermined database

object.

9. Apparatus according to claim 5, wherein said enforcement

rules each comprise at least one condition, said condition being the

specification, within the database operation text of any one of a group of

database objects belonging to a preformulated group.

10. Apparatus according to claim 1, further comprising a

management module for setting said enforcement rules.

11. Apparatus according to claim 10, wherein said management

module is further operable to define groups of database commands, said groups

being usable to define said rules.

12. Apparatus according to claim 10, wherein said management

module is accessible via a graphical user interface.

13. Apparatus according to claim 12, wherein said management

module is operable to set an access policy comprising a plurality of rules, each

rule specifying a group of commands, a group of users and a group of data

objects, said access policy being transferable to said rule application unit.

14. Apparatus according to claim 1, wherein at least some of said

enforcement rules define logging activities and said enforcement unit further

comprises logging functionality therefore.

15. Apparatus according to claim 1, wherein said inspecting

comprises obtaining a signature from said database operation text and wherein

at least some of said rules relate at least in part to said signatures.

16. Method for protection of database objects from unwanted access

comprising:

parsing a data packet to find database operation commands in said

packet,

applying enforcement rules to said data packet, said rules specifying

conditions based at least partially on said database operation commands,

thereby to protect respective database objects.

17. Method according to claim 16, further comprising finding

information regarding sources of respective data packets and regarding data

objects of respective commands.

18. Method according to claim 17, wherein said finding information

regarding sources of respective data packets is carried out per user connection,

the method further comprising associating said sources with data packets of

said user connection.

19. Method according to claim 17, further comprising applying said

enforcement rules selectively to said data packets, depending on information

found in respective data packets by said parsing units.

20. Method according to claim 19, wherein said database operation

commands are arranged in groups and wherein said rules relate to said groups.

21. Method according to claim 20, wherein said command sources

are arranged in groups and said rules relate to said groups.

22. Method according to claim 21, wherein said database objects are

arranged in groups and wherein said rules relate to said groups.

23. Method according to claim 16, further comprising arranging said

database commands into groups and using said groups to define said rules.

24. Method according to claim 23, further comprising forming an

access policy by arranging together a plurality of rules for use.