Anonymity and Security without compromise

About project

The go-peer library contains a large number of functions necessary to ensure the security of transmitted or stored information, as well as for the anonymity of nodes in the decentralized form. The library can be divided into several main modules:

1. The crypto module represents cryptographic primitives: 1) asymmetric encryption, decryption; 2) asymmetric signing and signature verification; 3) symmetric encryption and decryption; 4) hashing; 5) key building; 6) computational problems (puzzles); 7) cryptographically stable pseudorandom number generator.
2. The client module for encrypting and decrypting information with the attached data integrity (hash), authentication (signature) and confirmation (work). It is a basic part of the anonymity module.
3. The client/queue module represents the generation, storage and issuance of encrypted messages every time the period specified by the application is reached. Uses the client module.
4. The network module is a decentralized communication between network nodes. It does not represent any protection of information and anonymity of participants.
5. The network/anonymity module to ensure anonymity based on the fifth^ stage. Presents the main functions for working with the network on top of the network and queue modules.
6. The storage module includes crypto type storage. This type of storage can be used for secure storage of passwords and private keys.
7. The storage/database module is a key-value database with the functions of value encryption and key hashing inherited from the storage module.

Examples of works in the directory github.com/number571/go-peer/examples;

Installation

$ go get github.com/number571/go-peer

Requirements

1. Go version >= 1.16

Releases

All programs are compiled for {amd64, arm64} ARCH and {windows, linux, darwin} OS as pattern = appname_arch_os. In total, one application is compiled into six versions. The entire list of releases can be found here: github.com/number571/go-peer/releases.

Some final applications are compositions of other applications. Due to this, such applications are named in a slightly different way. Designation of abbreviations:

1. hlc - Hidden Lake Composite
2. hlc_sm - Composite of HLS (service) + HLM (messenger)
3. hlc_st - Composite of HLS + HLT (traffic)
4. hlc_stm - Composite of HLS + HLT + HLM

Dependencies

1. Go library (used by pkg/storage/database) github.com/syndtr/goleveldb;
2. Go library (used by cmd/hidden_lake/messenger) github.com/boombuler/barcode;
3. Go library (used by cmd/hidden_lake/messenger) golang.org/x/net;
4. Go library (used by pkg/crypto/keybuilder) golang.org/x/crypto;
5. CSS/JS library (used by cmd/hidden_lake/messenger) getbootstrap.com

Theoretical works

1. Theory of the structure of hidden systems
2. Monolithic cryptographic protocol
3. Abstract anonymous networks
4. Decentralized key exchange protocol

Also, the composition of these works can be found in the book The general theory of anonymous communications (Ridero). This book can be purchased in a tangible form on the Ozon and Wildberries marketplaces. You can download the book in digital form for free here.

Library based applications

1. Hidden Lake Service
2. Hidden Lake Messenger
3. Hidden Lake Traffic
4. Hidden Lake Adapters

1. Hidden Lake Service

github.com/number571/go-peer/cmd/hidden_lake/service

The Hidden Lake Service is the core of an anonymous network with theoretically provable anonymity. HLS is based on the fifth^ stage of anonymity and is an implementation of an abstract anonymous network based on queues. It is a peer-to-peer network communication with trusted friend-to-friend participants. All transmitted and received messages are in the form of end-to-end encryption.

Features / Anonymity networks	Queue-networks (5^stage)	Entropy-networks (6stage)	DC-networks (1^stage)
Theoretical provability	+	+	+
Ease of software implementation	+	-	-
Polymorphism of information	-	+	+
Static communication delay	+	-	+
Sending parallel messages	-	+	-
Network scales easily	-	-	-

A feature of HLS (compared to many other anonymous networks) is its easy adaptation to a hostile centralized environment. Anonymity can be restored literally from one node in the network, even if it is the only point of failure.

More information about HLS in the habr.com/ru/post/696504

How it works

Each network participant sets a message generation period for himself (the period can be a network constant for all system participants). When one cycle of the period ends and the next begins, each participant sends his encrypted message to all his connections (those in turn to all of their own, etc.). If there is no true message to send, then a pseudo message is generated (filled with random bytes) that looks like a normal encrypted one. The period property ensures the anonymity of the sender.

Figure 1. Queue and message generation in HLS.

Since the encrypted message does not disclose the recipient in any way, each network participant tries to decrypt the message with his private key. The true recipient is only the one who can decrypt the message. At the same time, the true recipient acts according to the protocol and further distributes the received packet, even knowing the meaninglessness of the subsequent dispatch. This property makes it impossible to determine the recipient.

Simple example of the client package (encrypt/decrypt functions) in the directory github.com/number571/go-peer/pkg/client/_examples;

Figure 2. Two participants are constantly generating messages for their periods on the network. It is impossible to determine their real activity.

Data exchange between network participants is carried out using application services. HLS has a dual role: 1) packages traffic from pure to anonymizing and vice versa; 2) converts external traffic to internal and vice versa. The second property is the redirection of traffic from the network to the local service and back.

Figure 3. Interaction of third-party services with the traffic anonymization service.

As shown in the figure above, HLS acts as an anonymizer and handlers of incoming and outgoing traffic. The remaining parts in the form of applications and services depend on third-party components (as an example, HLM).

Supported platforms

• Windows (x86_64, arm64)
• Linux (x86_64, arm64)
• MacOS (x86_64, arm64)

Minimum system requirements

1. Processor: 1x2.2GHz (limit of communication = ~10-15 nodes)
2. Memory: 0.5GB RAM (~250MB of memory is consumed)
3. Storage: 5Gib available space (the size of hashes per year from one node)

Build and run

Default build and run

$ cd ./cmd/hidden_lake/service
$ make build # create hls, hls_[arch=amd64,arm64]_[os=linux,windows,darwin] and copy to ./bin
$ make run # run ./bin/hls

> [INFO] 2023/06/03 14:32:40 HLS is running...
> [INFO] 2023/06/03 14:32:42 service=HLS type=BRDCS hash=43A5E9C5...BA73DF43 addr=211494E4...EEA12BBC proof=0000000002256145 conn=127.0.0.1:
> [INFO] 2023/06/03 14:32:47 service=HLS type=BRDCS hash=EFDDC1D4...C47588AD addr=211494E4...EEA12BBC proof=0000000000090086 conn=127.0.0.1:
> [INFO] 2023/06/03 14:32:52 service=HLS type=BRDCS hash=8549E257...EDEB2748 addr=211494E4...EEA12BBC proof=0000000000634328 conn=127.0.0.1:
> ...

Service was running with random private key. Open ports 9571 (TCP, traffic) and 9572 (HTTP, interface). Creates ./hls.cfg or ./_mounted/hls.cfg (docker) and ./hls.db or ./_mounted/hls.db (docker) files. The file hls.db stores hashes of sent/received messages.

Default config hls.cfg

{
	"settings": {
		"message_size_bytes": 8192,
		"work_size_bits": 20,
		"key_size_bits": 4096,
		"queue_period_ms": 5000,
		"limit_void_size_bytes": 4096,
		"messages_capacity": 2048
	},
	"logging": [
		"info",
		"warn",
		"erro"
	],
	"address": {
		"tcp": "127.0.0.1:9571",
		"http": "127.0.0.1:9572"
	},
	"services": {
		"go-peer/hidden-lake-messenger": "127.0.0.1:9592"
	}
}

If service works not in docker's environment than need rewrite connection host in hls.cfg file from messengerto IP address (example: 127.0.0.1:9592 for local network).

Build and run with docker

$ cd ./cmd/hidden_lake/service
$ make docker-build 
$ make docker-run

> [INFO] 2023/06/03 07:36:49 HLS is running...
> [INFO] 2023/06/03 07:36:51 service=HLS type=BRDCS hash=AF90439F...9F29A036 addr=BB58A8A2...B64D62C2 proof=0000000000479155 conn=127.0.0.1:
> [INFO] 2023/06/03 07:36:56 service=HLS type=BRDCS hash=2C4CE60A...E55BF9C4 addr=BB58A8A2...B64D62C2 proof=0000000000521434 conn=127.0.0.1:
> [INFO] 2023/06/03 07:37:01 service=HLS type=BRDCS hash=A9285F98...F96DB93D addr=BB58A8A2...B64D62C2 proof=0000000001256786 conn=127.0.0.1:
> ...

Example

There are three nodes in the network send_hls, recv_hls and middle_hls. The send_his and recv_hls nodes connects to middle_hls. As a result, a link of the form send_his <-> middle_hls <-> recv_hls is created. Due to the specifics of HLS, the centralized middle_hls node does not violate the security and anonymity of the send_hls and recv_hls subjects in any way. All nodes, including the middle_hls node, set periods and adhere to the protocol of constant message generation.

The recv_hls node contains its echo_service, which performs the role of redirecting the request body back to the client as a response. Access to this service is carried out by its alias hidden-echo-service, put forward by the recv_hls node.

...
// handle: "/echo"
// return format: {"echo":string,"return":int}
func echoPage(w http.ResponseWriter, r *http.Request) {
	if r.Method != "POST" {
		response(w, 2, "failed: incorrect method")
		return
	}
	res, err := io.ReadAll(r.Body)
	if err != nil {
		response(w, 3, "failed: read body")
		return
	}
	response(w, 1, string(res))
}
...

Identification between recv_hls and send_hls nodes is performed using public keys. This is the main method of identification and routing in the HLS network. IP addresses are only needed to connect to such a network and no more. Requests and responses structure are HEX encoded.

Structure of request. The body hello, world! is encoded base64.

JSON_DATA='{
        "method":"POST",
        "host":"hidden-echo-service",
        "path":"/echo",
        "head":{
            "Accept": "application/json"
        },
        "body":"aGVsbG8sIHdvcmxkIQ=="
}';

Request format

PUSH_FORMAT='{
        "receiver":"Alice",
        "hex_data":"'$(str2hex "$JSON_DATA")'"
}';

Build and run nodes

$ cd examples/echo_service/default
$ make

Logs from middle_hls node. When sending requests and receiving responses, middle_hls does not see the action. For him, all actions and moments of inaction are equivalent.

Figure 4. Output of all actions and all received traffic from the middle_hls node.

Send request

$ cd examples/echo_service
$ make request # go run ./_request/main.go
# OR
$ ./_request/request.sh

Get response

HTTP/1.1 200 OK
Date: Mon, 22 May 2023 18:18:34 GMT
Content-Length: 113
Content-Type: text/plain; charset=utf-8

{"code":200,"head":{"Content-Type":"application/json"},"body":"eyJlY2hvIjoiaGVsbG8sIHdvcmxkISIsInJldHVybiI6MX0K"}
Request took 8 seconds

Return code 200 is HTTP code = StatusOK. Decode base64 response body

echo "eyJlY2hvIjoiaGVsbG8sIHdvcmxkISIsInJldHVybiI6MX0K" | base64 -d
> {"echo":"hello, world!","return":1}

Figure 5. Example of running HLS with internal service.

Also you can run example with docker-compose. In this example, all nodes have logging enabled

$ cd examples/echo_service/_docker/default
$ make

Simple examples of the anonymity package in the directory github.com/number571/go-peer/pkg/network/anonymity/_examples;

2. Hidden Lake Messenger

github.com/number571/go-peer/cmd/hidden_lake/messenger;

The Hidden Lake Messenger is a messenger based on the core of an anonymous network with theoretically provable anonymity of HLS. A feature of this messenger is the provision of anonymity of the fact of transactions (sending, receiving).

HLM is an application that implements a graphical user interface (GUI) on a browser-based HTML/CSS/JS display. Most of the code is based on the bootstrap library https://getbootstrap.com /. GUI is adapted to the size of the window, so it can be used both in a desktop and in a smartphone.

More information about HLM in the habr.com/ru/post/701488

How it works

Most of the code is a call to API functions from the HLS kernel. Thanks to this approach, implicit authorization of users is formed from the state of the anonymizing service.

Figure 6. Example of chat room in HLM.

However, there are additional features aimed at the security of the HLM application itself. All messages are stored in a local database in encrypted form with a key formed from storage_key param.

Supported platforms

• Windows (x86_64, arm64)
• Linux (x86_64, arm64)
• MacOS (x86_64, arm64)

Build and run

Default build and run

$ cd ./cmd/hidden_lake/messenger
$ make build # create hlm, hlm_[arch=amd64,arm64]_[os=linux,windows,darwin] and copy to ./bin
$ make run # run ./bin/hlm

> [INFO] 2023/06/03 15:30:31 HLM is running...
> ...

Open ports 9591 (HTTP, interface) and 9592 (HTTP, incoming). Creates ./hlm.cfg or ./_mounted/hlm.cfg (docker) and ./hlm.db or ./_mounted/hlm.db (docker) files. The file hlm.db stores all sent/received messages in encrypted view.

Default config hlm.cfg

{
	"settings": {
		"messages_capacity": 2048
	},
	"logging": [
		"info",
		"warn",
		"erro"
	],
	"language": "ENG",
	"address": {
		"interface": "127.0.0.1:9591",
		"incoming": "127.0.0.1:9592"
	},
	"connection": "127.0.0.1:9572"
}

If messenger works not in docker's environment than need rewrite connection hosts in hlm.cfg file from service and traffic to IP addresses (example: 127.0.0.1:9572 and also 127.0.0.1:9581 for local network).

Build and run with docker

$ cd ./cmd/hidden_lake/messenger
$ make docker-build 
$ make docker-run

> [INFO] 2023/06/03 08:35:50 HLM is running...
> ...

Example

The example will involve (as well as in HLS) three nodes middle_hls, node1_hlm and node2_hlm. The first one is only needed for communication between node1_hlm and node2_hlm nodes. Each of the remaining ones is a combination of HLS and HLM, where HLM plays the role of an application and services, as it was depicted in Figure 3.

Build and run nodes

$ cd examples/anon_messenger/default
$ make

The output of the middle_hls node is similar to Figure 4. Than open browser on localhost:8080. It is a node1_hlm. This node is a Bob.

Figure 7. Home page of the HLM application.

To see the success of sending and receiving messages, you need to do all the same operations, but with localhost:7070 as node2_hlm. This node will be Alice.

Also you can run example with docker-compose. In this example, all nodes have logging enabled

$ cd examples/anon_messenger/_docker/default
$ make

Figure 8. Log of the three nodes with request/response actions.

More example images about HLM pages in the github.com/number571/go-peer/cmd/hidden_lake/messenger/_images

3. Hidden Lake Traffic

github.com/number571/go-peer/cmd/hidden_lake/traffic;

The Hidden Lake Traffic is an application that saves traffic passed through HLS. The saved traffic can be used by other applications when they were offline. HLT provides an API for loading and unloading messages. Messages are stored in the database based on the "ring" structure. Thus, new messages will overwrite the old ones after some time.

More information about HLT in the habr.com/ru/post/717184

How it works

HLT emulates HLS to receive messages. In this scenario, HLT has only the functions of accepting messages, without the ability to generate or send them via HLS or independently.

Figure 9. Example of running HLT client.

Supported platforms

• Windows (x86_64, arm64)
• Linux (x86_64, arm64)
• MacOS (x86_64, arm64)

Minimum system requirements

1. Processor: 1x2.2GHz (characteristic for relayer mode)
2. Memory: 0.5GB RAM (~200MB of memory is consumed in relayer mode)
3. Storage: 20Gib available space (the size of messages per day from one node in storage mode)

Build and run

Default build and run

$ cd ./cmd/hidden_lake/traffic
$ make build # create hlt, hlt_[arch=amd64,arm64]_[os=linux,windows,darwin] and copy to ./bin
$ make run # run ./bin/hlt

> [INFO] 2023/06/03 15:39:13 HLT is running...
> ...

Open ports 9581 (HTTP, interface). Creates ./hlt.cfg or ./_mounted/hlt.cfg (docker), ./hlt.db or ./_mounted/hlt.db (docker) files. The file hlm.db stores all sent/received messages as structure ring from network HL.

Default config hlt.cfg

{
	"settings": {
		"message_size_bytes": 8192,
		"work_size_bits": 20,
		"messages_capacity": 2048,
		"queue_period_ms": 5000,
		"limit_void_size_bytes": 4096
	},
	"logging": [
		"info",
		"warn",
		"erro"
	],
	"address": {
		"tcp": "127.0.0.1:9581",
		"http": "127.0.0.1:9582"
	},
	"connections": [
		"127.0.0.1:9571"
	]
}

If traffic works not in docker's environment than need rewrite connection host in hlt.cfg file from service to IP address (example: 127.0.0.1:9571 for local network).

Build and run with docker

$ cd ./cmd/hidden_lake/traffic
$ make docker-build 
$ make docker-run

> [INFO] 2023/06/03 08:44:14 HLT is running...
> ...

Example

Build and run service

$ cd examples/traffic_keeper
$ make

Run client

$ cd client
$ go run ./main.go w 'hello, world!'
$ go run ./main.go h
$ go run ./main.go r cb3c6558fe0cb64d0d2bad42dffc0f0d9b0f144bc24bb8f2ba06313af9297be4 # hash get by 'h' option

4. Hidden Lake Adapters

github.com/number571/go-peer/cmd/hidden_lake/adapters

The Hidden Lake Adapters are a way to exchange data between multiple HLS processes via third-party services. Thus, there is no need to use your own computing resources in the face of individual servers to store or distribute the traffic generated by HLS.

More information about HLA in the habr.com/ru/post/720544

How it works

Adapters in their full execution represent one design template - "Flyweight". They are based on the composition of HLS and HLT processes. Adapters adapt to the interfaces of the service for reading/writing data and, thanks to this, are able to conduct anonymizing traffic through the service.

Figure 10. Architecture of HLA.

Example

Since the adapters are capable of performing the same functions as the source programs, without secret communication channels, then all the examples become similar. So, let's analyze one of the examples based on the echo service.

Build and run nodes

$ cd examples/echo_service/secret_channel
$ make

Send request

$ cd examples/echo_service
$ ./request.sh

Get response

HTTP/1.1 200 OK
Date: Mon, 22 May 2023 18:18:34 GMT
Content-Length: 113
Content-Type: text/plain; charset=utf-8

{"code":200,"head":{"Content-Type":"application/json"},"body":"eyJlY2hvIjoiaGVsbG8sIHdvcmxkISIsInJldHVybiI6MX0K"}
Request took 8 seconds

There are no external differences, but there are internal ones. While the original model assumed the presence of a middle_hls node through which all traffic was broadcast, there is no such intermediate node in the model based on secret communication channels, there is a service that performs its own logical functions that are in no way tied to traffic anonymization. And, thus, adapters use a third-party service in order to pass traffic through it.

Figure 11. Example of running HLA client.

Similarly, you can use a more complex composition, as shown in the example examples/anon_messenger/secret_channel.

Deprecated applications

• Hidden Lake (can be used as HLS+HLM+HLT+...): github.com/number571/hidden-lake;
• Hidden Email Service (HLT can be used as HES): github.com/number571/hes;
• Hidden Lake Service (new release = HLS): github.com/number571/hls;

License

Licensed under the MIT License. See LICENSE for the full license text.