How to protect Python source from modification

[Frank Millman]

Hi all

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program. As it is written in Python, with source available, this would
be quite easy. My target market extends well up into the mid-range, but
I do not think that any CFO would contemplate using a program that is
so open to manipulation.

The only truly secure solution I can think of would involve a radical
reorganisation of my program, so I am writing to see if anyone has a
simpler suggestion. Here is the idea.

1. Write a socket server program that runs on the server. The socket
server is the only program that connects to the database. The client
program connects to the server, which authenticates the client against
the database and then listens for requests from the client.

2. Devise my own protocol for communication between client and server.
For selects, the client sends a request, the server checks permissions,
then retrieves the data from the database and passes it to the client.
For updates, the client passes up the data to be updated, the server
checks it against its business logic, and then updates the database.

There is the question of where state should be maintained. If on the
server, I would have to keep all the client/server connections open,
and maintain the state of all the sessions, which would put quite a
load on the server. If on the client, I would have to reorganise my
thinking even more, but this would have an advantage - I will
eventually want to write a browser interface, and this would be much
closer in concept, so the two approaches would be quite similar.

This raises the question of whether I should even bother with a gui
client, or bite the bullet and only have a browser based front end.
Judging from recent comments about new technologies such as Ajax, a lot
of the disadvantages have been overcome, so maybe this is the way to
go.

It would be a shame to scrap all the effort I have put into my
wxPython-based front end. On the other hand, it would be pointless to
continue with an approach that is never going to give me what I want.
Any advice which helps to clarify my thinking will be much appreciated.

Thanks

Frank Millman

[Gerhard Häring]

Frank Millman wrote:

Hi all

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program. As it is written in Python, with source available, this would
be quite easy. My target market extends well up into the mid-range, but
I do not think that any CFO would contemplate using a program that is
so open to manipulation. [...]

My suggestion is to use py2exe or cx_Freeze to package your application.
It's then not as trivial to modify it. Btw. you don't need to ship the
..py source code files, it's enough to ship only .pyc bytecode files.

Using py2exe it's not even obvious that your application is written in
Python at all.

It's not a silver bullet, but at least it makes recompiling/modifiying
your app not easier than with Java (and/or .NET I suppose).

That being said, even if you continue with the GUI approach, it may
still be a good idea to factor out all the business logic in a separate
module so you can eventually switch to a web application or a three-tier
model without too much effort.

Also, there's no need at all to put in countless hours implementing your
own network protocol. If you really want to separate client and app
server, then why not use something simple as PyRO, or even XML/RPC.

HTH,

-- Gerhard

[Peter Hansen]

Frank Millman wrote:

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program. As it is written in Python, with source available, this would
be quite easy. My target market extends well up into the mid-range, but
I do not think that any CFO would contemplate using a program that is
so open to manipulation.

The only truly secure solution I can think of would involve a radical
reorganisation of my program

Please define what "truly secure" means to you.

I think you'll find that the only "truly secure" solution is to install
the critical authentication and business logic stuff that you want to
protect on a server to which the user does not have physical access.

People wanting to protect critical algorithms often conclude that they
need to have a "black box" server which cannot be physically opened by
the user.

Or do you think this issue is in some way unique to Python? You might
not realize that the only difference from a security point of view
between shipping such a program written in Python and one written in,
say, C++, is that "modifying the program" is somewhat more tedious with
C++. That's no better than security by obscurity; maybe it should be
called "security by adiposity". ;-)

But the real answer does depend a lot on *exactly* what kind of security
you want (or, ultimately, what it turns out you really need, once you've
clarified your thinking based on the feedback you do get here). Issues
like: are you more concerned about detecting changes, or in preventing
them in the first place? (the latter is much harder); what is the nature
of software that competes with yours? (is it really any more secure, or
only apparently so? maybe this is just a marketing issue); and is there
any intellectual property that you are trying to protect here, or are
you just interested in avoiding casual disruption of normal operation?

-Peter

[Frank Millman]

Gerhard Häring wrote:

Frank Millman wrote:

Hi all

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program. As it is written in Python, with source available, this would
be quite easy. My target market extends well up into the mid-range, but
I do not think that any CFO would contemplate using a program that is
so open to manipulation. [...]
My suggestion is to use py2exe or cx_Freeze to package your application.
It's then not as trivial to modify it. Btw. you don't need to ship the
.py source code files, it's enough to ship only .pyc bytecode files.

Using py2exe it's not even obvious that your application is written in
Python at all.

It's not a silver bullet, but at least it makes recompiling/modifiying
your app not easier than with Java (and/or .NET I suppose).

My problem is that, if someone has access to the network and to a
Python interpreter, they can get hold of a copy of my program and use
it to knock up their own client program that makes a connection to the
database. They can then execute any arbitrary SQL command.
That being said, even if you continue with the GUI approach, it may
still be a good idea to factor out all the business logic in a separate
module so you can eventually switch to a web application or a three-tier
model without too much effort.

Agreed
Also, there's no need at all to put in countless hours implementing your
own network protocol. If you really want to separate client and app
server, then why not use something simple as PyRO, or even XML/RPC.

Perhaps 'protocol' is the wrong word. I already have a simple socket
server program running. If explain how I do it, perhaps you can
indicate whether PyRO or XML/RPC would make my life easier.

The server program is currently programmed to accept a number of
message types from the client program. Each message's data string
starts with a numeric prefix, which indicates the type of message,
followed by a pickled tuple of arguments. The server program reads the
string, extracts the numeric prefix, and passes the rest of the string
to the appropriate function using a subthread.

For example, I keep track of who is currently logged in. On startup,
the client connects to my server and sends a '1' followed by their
userid and other information. The server receives this and passed the
data to a 'login' function, which uses a Python dictionary to store the
information. If the server detects that the user is already logged in,
it sends back an error code and the client program displays a message
and terminates. Otherwise it sends back an 'ok' code, and the client
can continue. When the client logs off, it sends a '2' followed by
their userid, which the server receives and passes it to a 'logoff'
function, which deletes the entry from the dictionary.

The system of numeric prefixes and associated data string making up a
message is what I mean by a protocol.
HTH,

-- Gerhard

Thanks

Frank

[bruno modulix]

Frank Millman wrote:

Hi all

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program.
If your program relies on a RDBMS, then it's the RDBMS job to enforce
security rules.
As it is written in Python, with source available, this would
be quite easy.

Then there's probably something wrong with the way you manage security.

NB: splitting business logic from the GUI is still a good idea anyway.

--
bruno desthuilliers - unpythonic sig:
python -c "print '@'.join(['.'.join([w[::-1] for w in p.split('.')]) for
p in 'o****@xiludom. gro'.split('@')])"

[Frank Millman]

Peter Hansen wrote:

Frank Millman wrote:

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program. As it is written in Python, with source available, this would
be quite easy. My target market extends well up into the mid-range, but
I do not think that any CFO would contemplate using a program that is
so open to manipulation.

The only truly secure solution I can think of would involve a radical
reorganisation of my program
Please define what "truly secure" means to you.

Fair question. I am not expecting 'truly' to mean 100% - I know that is
impossible. I will try to explain.

Here are some assumptions -
1. A system adminstrator is responsible for the system.
2. There is a single userid and password for connecting to the
database. This must be stored somewhere so that the client program can
read it to generate the appropriate connection string. The users do not
need to know this userid and password.
3. Each user has their own userid and password, which is stored in the
database in a 'users' table. I use this in my program for
authentication when a user tries to connect.
4. The client program can be run from anywhere on the network that has
access to the program and to a Python interpreter.

[snip]

But the real answer does depend a lot on *exactly* what kind of security
you want (or, ultimately, what it turns out you really need, once you've
clarified your thinking based on the feedback you do get here). Issues
like: are you more concerned about detecting changes, or in preventing
them in the first place? (the latter is much harder); what is the nature
of software that competes with yours? (is it really any more secure, or
only apparently so? maybe this is just a marketing issue); and is there
any intellectual property that you are trying to protect here, or are
you just interested in avoiding casual disruption of normal operation?

I am not concerned about anyone reading my code - in fact I am looking
forward to releasing the source and getting some feedback.

My concern is this. I have all this fancy authentication and business
logic in my program. If someone wants to bypass this and get direct
access to the database, it seems trivially easy. All they have to do is
read my source, find out where I get the connection string from, write
their own program to make a connection to the database, and execute any
SQL command they want.

If I move all the authentication and business logic to a program which
runs on the server, it is up to the system administrator to ensure that
only authorised people have read/write/execute privileges on that
program. Clients will have no privileges, not even execute. They will
have their own client program, which has to connect to my server
program, and communicate with it in predefined ways. I *think* that in
this way I can ensure that they cannot do anything outside the bounds
of what I allow them.

The only problem is that this is very different from the way my program
works at present, so it will be quite a bit of work to re-engineer it.
If someone can suggest a simpler solution obviously I would prefer it.
But if the consensus is that I am thinking along the right lines, I
will roll up my sleeves and get stuck in.
-Peter

I hope this explains my thinking a bit better.

Thanks

Frank

[bruno modulix]

Frank Millman wrote:

Peter Hansen wrote:

Frank Millman wrote:
(snip)

The only truly secure solution I can think of would involve a radical
reorganisati on of my program
Please define what "truly secure" means to you.

Fair question. I am not expecting 'truly' to mean 100% - I know that is
impossible. I will try to explain.

Here are some assumptions -
1. A system adminstrator is responsible for the system.
2. There is a single userid and password for connecting to the
database. This must be stored somewhere so that the client program can
read it to generate the appropriate connection string. The users do not
need to know this userid and password.
3. Each user has their own userid and password,
which is stored in the
database in a 'users' table. I use this in my program for
authentication when a user tries to connect.

Why not simply using the security system of your RDBMS ? If you set up
appropriate privileges in the RDBMS, you won't have to store any
userid/password in the program, and no user will be able to bypass
anything, even if connecting directly (like with a CLI DB client) to the
RDBMS.

[snip]

(snip more)
I am not concerned about anyone reading my code - in fact I am looking
forward to releasing the source and getting some feedback.

My concern is this. I have all this fancy authentication and business
logic in my program. If someone wants to bypass this and get direct
access to the database, it seems trivially easy. All they have to do is
read my source, find out where I get the connection string from, write
their own program to make a connection to the database, and execute any
SQL command they want.

That's why RDBMS have an authentication and security system. This
doesn't means your program doesn't have or cannot add it's own security
management, but it should be based on the RDBMS one.

--
bruno desthuilliers
python -c "print '@'.join(['.'.join([w[::-1] for w in p.split('.')]) for
p in 'o****@xiludom. gro'.split('@')])"

[Dennis Lee Bieber]

On 12 Sep 2005 08:33:10 -0700, "Frank Millman" <fr***@chagford .com>
declaimed the following in comp.lang.pytho n:

My problem is that, if someone has access to the network and to a
Python interpreter, they can get hold of a copy of my program and use
it to knock up their own client program that makes a connection to the
database. They can then execute any arbitrary SQL command.
If your DBMS is directly accessible on the net, you're vulnerable
even without Python. Especially if you have "authentication " logic being
done at the client end. There is nothing to prevent someone using a
compatible query browser or command-line utility to make connection
attempts to the server, followed by classical username/password cracking
stuff.

The server program is currently programmed to accept a number of
message types from the client program. Each message's data string
starts with a numeric prefix, which indicates the type of message,
followed by a pickled tuple of arguments. The server program reads the
string, extracts the numeric prefix, and passes the rest of the string
to the appropriate function using a subthread.
Ah, okay -- you /do/ already have something running in the middle.
For example, I keep track of who is currently logged in. On startup,
the client connects to my server and sends a '1' followed by their
userid and other information. The server receives this and passed the
data to a 'login' function, which uses a Python dictionary to store the
information. If the server detects that the user is already logged in,
it sends back an error code and the client program displays a message
and terminates. Otherwise it sends back an 'ok' code, and the client
can continue. When the client logs off, it sends a '2' followed by
their userid, which the server receives and passes it to a 'logoff'
function, which deletes the entry from the dictionary.
Obscuring the Python stuff will only be a minor delay factor in
breaking that -- someone really serious could probably stick in a packet
sniffer and record a transaction sequence, eventually reverse mapping
back to the types of operations each code represents.

Database security? First step would be to USE the DBMS privilege
system to limit operations to only those SQL statements, tables, and
data columns that are needed for your client program; since you appear
to be using user/password information already, each such user could have
different privileges, limiting some to retrieval only, for example. As
for your "server", I'd probably start a thread for each connected user,
so that thread handles all communication. Your description sounds more
like a rudimentary proxy adding in a counting scheme, but not really
isolating separate client connections.
-- =============== =============== =============== =============== == <
wl*****@ix.netc om.com | Wulfraed Dennis Lee Bieber KD6MOG <
wu******@dm.net | Bestiaria Support Staff <
=============== =============== =============== =============== == <
Home Page: <http://www.dm.net/~wulfraed/> <
Overflow Page: <http://wlfraed.home.ne tcom.com/> <

[Frank Millman]

bruno modulix wrote:

Frank Millman wrote:

Hi all

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

The client program contains all the authentication and business logic.
It has dawned on me that anyone can bypass this by modifying the
program.
If your program relies on a RDBMS, then it's the RDBMS job to enforce
security rules.

Two possible responses to this -

1. You are right (90% probability)

2. I have certain requirements which can not easily be expressed in the
RDBMS, so it is easier to use the application to enforce certain rules
(10% probability)

Unfortunately I am stuck with number 2 at present.

As it is written in Python, with source available, this would
be quite easy.

Then there's probably something wrong with the way you manage security.

Probably - I am learning the hard way <g>
NB: splitting business logic from the GUI is still a good idea anyway.

I do have it fairly well split, but it all ends up being processed on
the client, which I think is the root of my problem.
--
bruno desthuilliers - unpythonic sig:
python -c "print '@'.join(['.'.join([w[::-1] for w in p.split('.')]) for
p in 'o****@xiludom. gro'.split('@')])"

Thanks

Frank

[Bugs]

As a side question Frank, how was your experiences using wxPython for
your GUI?
Any regrets choosing wxPyton over another toolkit?
Was it very buggy?
How was it to work with in general?
Any other real-world wxPython feedback you have is appreciated.

Frank Millman wrote:

I am writing a multi-user accounting/business system. Data is stored in
a database (PostgreSQL on Linux, SQL Server on Windows). I have written
a Python program to run on the client, which uses wxPython as a gui,
and connects to the database via TCP/IP.

<snip>