Connecting Tech Pros Worldwide Forums | Help | Site Map

Code protection

Boni
Guest
 
Posts: n/a
#1: Nov 17 '05
Dear all,

in order to protect my assembly component from decompilation I implemented
following schema:

I created mixed mode C++ project wich has managed cProxy class and unmanaged
cMemLoader. In the cMemLoader I load encrypted assembly, decrypt it into
memory buffer and then marshal this buffer into managed memory and load the
assembly from memory.

Although I am not sure if this approach is good enough and how easy is it to
break.

Can anybody comment on this? Improvements are welcome too.

Thanks in advance,

Boni



void cProxy::Load(void){

m_pLoader=new cMemLoader(C:\\edit\\MyEncriptedManaged.dll

m_pLoader->DecriptLoad();


unsigned char _pManagedBuff __gc[]=__gc new unsigned char __gc
[m_pLoader->m_file_size] ;

Marshal::Copy(m_pLoader->m_pBuff,_pManagedBuff,0,m_pLoader->m_file_size);

delete m_pLoader;

m_Asm= Assembly::Load(_pManagedBuff);

System::Type *_Types __gc[]= m_Asm->GetTypes();

Object *_Instance = 0;

MethodInfo *_Methods __gc[]= _Types[0]->GetMethods();

_Instance = System::Activator::CreateInstance(_Types[0]);

_Methods[0]->Invoke(_Instance,0);

}




Ignacio Machin \( .NET/ C# MVP \)
Guest
 
Posts: n/a
#2: Nov 17 '05

re: Code protection


Hi,

Did you try to obfuscate it first?

I would suggest you to use the tools accesible to the framework. In your
solution you encrypt the dll, meaning that the file in the HDD is not longer
a DLL, it's a data file. hence the framework cannot do nothing with it, you
cannot register it in the GAC, not using versionsing.

I would give a try to obfuscation first. Take a look at the community
version shipped with .NET


cheers,

--
Ignacio Machin,
ignacio.machin AT dot.state.fl.us
Florida Department Of Transportation


"Boni" <oilia@nospam> wrote in message
news:OSuiEWtOFHA.576@TK2MSFTNGP15.phx.gbl...[color=blue]
> Dear all,
>
> in order to protect my assembly component from decompilation I implemented
> following schema:
>
> I created mixed mode C++ project wich has managed cProxy class and
> unmanaged cMemLoader. In the cMemLoader I load encrypted assembly, decrypt
> it into memory buffer and then marshal this buffer into managed memory and
> load the assembly from memory.
>
> Although I am not sure if this approach is good enough and how easy is it
> to break.
>
> Can anybody comment on this? Improvements are welcome too.
>
> Thanks in advance,
>
> Boni
>
>
>
> void cProxy::Load(void){
>
> m_pLoader=new cMemLoader(C:\\edit\\MyEncriptedManaged.dll
>
> m_pLoader->DecriptLoad();
>
>
> unsigned char _pManagedBuff __gc[]=__gc new unsigned char __gc
> [m_pLoader->m_file_size] ;
>
> Marshal::Copy(m_pLoader->m_pBuff,_pManagedBuff,0,m_pLoader->m_file_size);
>
> delete m_pLoader;
>
> m_Asm= Assembly::Load(_pManagedBuff);
>
> System::Type *_Types __gc[]= m_Asm->GetTypes();
>
> Object *_Instance = 0;
>
> MethodInfo *_Methods __gc[]= _Types[0]->GetMethods();
>
> _Instance = System::Activator::CreateInstance(_Types[0]);
>
> _Methods[0]->Invoke(_Instance,0);
>
> }
>
>
>
>[/color]


Boni
Guest
 
Posts: n/a
#3: Nov 17 '05

re: Code protection


Dear Ignacio, dear all,

of course the dll is obfuscated before encryption. But just obfuscation is
in my case definitely not enougth protection (and it is the point, where for
me the decision is already done and will not be changed!!!). Please don't
misunderstand me, I don't want to discuss if it is worth to encrypt, but HOW
to encrypt so that the dll can't be decompiled.

Please could anybody comment on the strength of my approach or suggest how
to improve it?

Thanks so much,

Boni



"Ignacio Machin ( .NET/ C# MVP )" <ignacio.machin AT dot.state.fl.us>
schrieb im Newsbeitrag news:OMmhprtOFHA.3380@TK2MSFTNGP15.phx.gbl...[color=blue]
> Hi,
>
> Did you try to obfuscate it first?
>
> I would suggest you to use the tools accesible to the framework. In your
> solution you encrypt the dll, meaning that the file in the HDD is not
> longer a DLL, it's a data file. hence the framework cannot do nothing
> with it, you cannot register it in the GAC, not using versionsing.
>
> I would give a try to obfuscation first. Take a look at the community
> version shipped with .NET
>
>
> cheers,
>
> --
> Ignacio Machin,
> ignacio.machin AT dot.state.fl.us
> Florida Department Of Transportation
>
>
> "Boni" <oilia@nospam> wrote in message
> news:OSuiEWtOFHA.576@TK2MSFTNGP15.phx.gbl...[color=green]
>> Dear all,
>>
>> in order to protect my assembly component from decompilation I
>> implemented following schema:
>>
>> I created mixed mode C++ project wich has managed cProxy class and
>> unmanaged cMemLoader. In the cMemLoader I load encrypted assembly,
>> decrypt it into memory buffer and then marshal this buffer into managed
>> memory and load the assembly from memory.
>>
>> Although I am not sure if this approach is good enough and how easy is it
>> to break.
>>
>> Can anybody comment on this? Improvements are welcome too.
>>
>> Thanks in advance,
>>
>> Boni
>>
>>
>>
>> void cProxy::Load(void){
>>
>> m_pLoader=new cMemLoader(C:\\edit\\MyEncriptedManaged.dll
>>
>> m_pLoader->DecriptLoad();
>>
>>
>> unsigned char _pManagedBuff __gc[]=__gc new unsigned char __gc
>> [m_pLoader->m_file_size] ;
>>
>> Marshal::Copy(m_pLoader->m_pBuff,_pManagedBuff,0,m_pLoader->m_file_size);
>>
>> delete m_pLoader;
>>
>> m_Asm= Assembly::Load(_pManagedBuff);
>>
>> System::Type *_Types __gc[]= m_Asm->GetTypes();
>>
>> Object *_Instance = 0;
>>
>> MethodInfo *_Methods __gc[]= _Types[0]->GetMethods();
>>
>> _Instance = System::Activator::CreateInstance(_Types[0]);
>>
>> _Methods[0]->Invoke(_Instance,0);
>>
>> }
>>
>>
>>
>>[/color]
>
>[/color]


Mattias Sjögren
Guest
 
Posts: n/a
#4: Nov 17 '05

re: Code protection


>but HOW to encrypt so that the dll can't be decompiled.

If your code can do it, so can I.

[color=blue]
>Please could anybody comment on the strength of my approach or suggest how
>to improve it?[/color]

I would say it's pretty useless. It's trivial to break your code at
the right place and dump the content of _pManagedBuff.



Mattias

--
Mattias Sjögren [MVP] mattias @ mvps.org
http://www.msjogren.net/dotnet/ | http://www.dotnetinterop.com
Please reply only to the newsgroup.
Boni
Guest
 
Posts: n/a
#5: Nov 17 '05

re: Code protection


Hi Mattias,
could you suggest how to improve the approach, that it is not so trivial to
break?
Thanks,
Boni
"Mattias Sjögren" <mattias.dont.want.spam@mvps.org> schrieb im Newsbeitrag
news:ukXr4MuOFHA.3560@TK2MSFTNGP14.phx.gbl...[color=blue][color=green]
> >but HOW to encrypt so that the dll can't be decompiled.[/color]
>
> If your code can do it, so can I.
>
>[color=green]
>>Please could anybody comment on the strength of my approach or suggest how
>>to improve it?[/color]
>
> I would say it's pretty useless. It's trivial to break your code at
> the right place and dump the content of _pManagedBuff.
>
>
>
> Mattias
>
> --
> Mattias Sjögren [MVP] mattias @ mvps.org
> http://www.msjogren.net/dotnet/ | http://www.dotnetinterop.com
> Please reply only to the newsgroup.[/color]


Bill
Guest
 
Posts: n/a
#6: Nov 17 '05

re: Code protection


Are you trying to protect against proprietary algorithms or are you trying
to prevent someone from thwarting your licensing system? You have to way
your options... Is spending all this extra time attempting to write an
"unbreakable" scheme worth more to you then putting that time and effort in
to marketing your application? You could waste your entire life away
trying to protect your software.

Even if you did find the perfect protection what will happen is that someone
will use a stolen credit card to order the software and get it for free
anyway.



"Boni" <oilia@nospam> wrote in message
news:OoYJCWuOFHA.4028@TK2MSFTNGP10.phx.gbl...[color=blue]
> Hi Mattias,
> could you suggest how to improve the approach, that it is not so trivial
> to break?
> Thanks,
> Boni
> "Mattias Sjögren" <mattias.dont.want.spam@mvps.org> schrieb im Newsbeitrag
> news:ukXr4MuOFHA.3560@TK2MSFTNGP14.phx.gbl...[color=green][color=darkred]
>> >but HOW to encrypt so that the dll can't be decompiled.[/color]
>>
>> If your code can do it, so can I.
>>
>>[color=darkred]
>>>Please could anybody comment on the strength of my approach or suggest
>>>how
>>>to improve it?[/color]
>>
>> I would say it's pretty useless. It's trivial to break your code at
>> the right place and dump the content of _pManagedBuff.
>>
>>
>>
>> Mattias
>>
>> --
>> Mattias Sjögren [MVP] mattias @ mvps.org
>> http://www.msjogren.net/dotnet/ | http://www.dotnetinterop.com
>> Please reply only to the newsgroup.[/color]
>
>[/color]


Boni
Guest
 
Posts: n/a
#7: Nov 17 '05

re: Code protection


I am trying to protect algorithms written in managed code. By the way the
license manager runs in the unmanaged code, and if lisence is not valid
assembly will not be decripted.
So my only need is to prevent easy decompile in order to protect algorithms.
May be somebody has an idea how to improve my protection approach?
Thanks,

P.S. Please, I know not anything can be broken. I am trying just to raise
the level.
"Bill" <msgdev@hotmail.com> schrieb im Newsbeitrag
news:%23argCjuOFHA.1932@tk2msftngp13.phx.gbl...[color=blue]
> Are you trying to protect against proprietary algorithms or are you trying
> to prevent someone from thwarting your licensing system? You have to way
> your options... Is spending all this extra time attempting to write an
> "unbreakable" scheme worth more to you then putting that time and effort
> in to marketing your application? You could waste your entire life away
> trying to protect your software.
>
> Even if you did find the perfect protection what will happen is that
> someone will use a stolen credit card to order the software and get it for
> free anyway.
>
>
>
> "Boni" <oilia@nospam> wrote in message
> news:OoYJCWuOFHA.4028@TK2MSFTNGP10.phx.gbl...[color=green]
>> Hi Mattias,
>> could you suggest how to improve the approach, that it is not so trivial
>> to break?
>> Thanks,
>> Boni
>> "Mattias Sjögren" <mattias.dont.want.spam@mvps.org> schrieb im
>> Newsbeitrag news:ukXr4MuOFHA.3560@TK2MSFTNGP14.phx.gbl...[color=darkred]
>>> >but HOW to encrypt so that the dll can't be decompiled.
>>>
>>> If your code can do it, so can I.
>>>
>>>
>>>>Please could anybody comment on the strength of my approach or suggest
>>>>how
>>>>to improve it?
>>>
>>> I would say it's pretty useless. It's trivial to break your code at
>>> the right place and dump the content of _pManagedBuff.
>>>
>>>
>>>
>>> Mattias
>>>
>>> --
>>> Mattias Sjögren [MVP] mattias @ mvps.org
>>> http://www.msjogren.net/dotnet/ | http://www.dotnetinterop.com
>>> Please reply only to the newsgroup.[/color]
>>
>>[/color]
>
>[/color]


Bob Grommes
Guest
 
Posts: n/a
#8: Nov 17 '05

re: Code protection


I think your approach (obfuscation plus encryption) probably makes it as
hard as it's practical to make it -- maybe even harder. But neither you nor
your employer should labor under the misconception that there's any such
thing as bullet-proof protection. It's exactly as you put it in an earlier
message -- you are making it "not so trivial to break" -- not "impossible to
break". So long as you understand the distinction, well and good.

But in my experience, it's still better to innovate your competition into
irrelevance than to expend much serious effort in protection schemes. After
all if someone steals the code all they have is unsupported warez. I'm sure
you add enough value that any serious user would gladly pay for the product.
If not, you need to lower your price or add more value, or both. IMO,
protection schemes should only stop the most casual theives and keep honest
people honest, no more than that.

--Bob

"Boni" <oilia@nospam> wrote in message
news:uD4yb%23uOFHA.164@TK2MSFTNGP12.phx.gbl...[color=blue]
>I am trying to protect algorithms written in managed code. By the way the
>license manager runs in the unmanaged code, and if lisence is not valid
>assembly will not be decripted.
> So my only need is to prevent easy decompile in order to protect
> algorithms. May be somebody has an idea how to improve my protection
> approach?
> Thanks,
>
> P.S. Please, I know not anything can be broken. I am trying just to raise
> the level.
> "Bill" <msgdev@hotmail.com> schrieb im Newsbeitrag
> news:%23argCjuOFHA.1932@tk2msftngp13.phx.gbl...[color=green]
>> Are you trying to protect against proprietary algorithms or are you
>> trying to prevent someone from thwarting your licensing system? You have
>> to way your options... Is spending all this extra time attempting to
>> write an "unbreakable" scheme worth more to you then putting that time
>> and effort in to marketing your application? You could waste your
>> entire life away trying to protect your software.
>>
>> Even if you did find the perfect protection what will happen is that
>> someone will use a stolen credit card to order the software and get it
>> for free anyway.
>>
>>
>>
>> "Boni" <oilia@nospam> wrote in message
>> news:OoYJCWuOFHA.4028@TK2MSFTNGP10.phx.gbl...[color=darkred]
>>> Hi Mattias,
>>> could you suggest how to improve the approach, that it is not so trivial
>>> to break?
>>> Thanks,
>>> Boni
>>> "Mattias Sjögren" <mattias.dont.want.spam@mvps.org> schrieb im
>>> Newsbeitrag news:ukXr4MuOFHA.3560@TK2MSFTNGP14.phx.gbl...
>>>> >but HOW to encrypt so that the dll can't be decompiled.
>>>>
>>>> If your code can do it, so can I.
>>>>
>>>>
>>>>>Please could anybody comment on the strength of my approach or suggest
>>>>>how
>>>>>to improve it?
>>>>
>>>> I would say it's pretty useless. It's trivial to break your code at
>>>> the right place and dump the content of _pManagedBuff.
>>>>
>>>>
>>>>
>>>> Mattias
>>>>
>>>> --
>>>> Mattias Sjögren [MVP] mattias @ mvps.org
>>>> http://www.msjogren.net/dotnet/ | http://www.dotnetinterop.com
>>>> Please reply only to the newsgroup.
>>>
>>>[/color]
>>
>>[/color]
>
>[/color]


Closed Thread