Faisal's Notebook

Recent Notes

Process Management Journal Review 2025

6 min read

Sources

Use for significance of the study

  1. Y. Judith Petrizia, R. Sujitha, Sree Saradha M, International Journal of Advanced Research in Science, Communication and Technology, “Overview of Process Management in Operating Systems” - 2025-02
  2. Buerian Soongpol, Paniti Netinant, Meennapa Rukhiran, Sustainability, “Practical Sustainable Software Development in Architectural Flexibility for Energy Efficiency Using the Extended Agile Framework” - 2024
  3. Amin Vahdat, Alvin Lebeck, Carla Schlatter Ellis, Proceedings of the 9th workshop on ACM SIGOPS European workshop: beyond the PC: new challenges for the operating system, “Every Joule is Precious: The Case for Revisiting Operating System Design for Energy Efficiency” - 2000-10-17

Use for related literature and issues. REFERENCE SHOULD BE <5 YEARS OLD

  1. Hamed Dinari, International Journal of Modern Education and Computer Science, “Inter-Process Communication (IPC) in Distributed Environments: An Investigation and Performance Analysis of Some Middleware Technologies” - 2020-04-08
  2. Christos Lyvas, Costas Lambrinoudakis, Dimitris Geneiatakis, International Journal of Information Security, “IntentAuth: Securing Android’s Intent-based inter-process communication” - 2022-04-22
  3. Dinglan Peng, Congyu Liu, Tapti Palit, Pedro Fonseca, Anjo Vahldiek-Oberwagner, Mona Vij, IEEE, “μSwitch: Fast Kernel Context Isolation with Implicit Context Switches” - 2023-07-21
  4. Jack Tigar Humphries, Kostis Kaffes, David Mazières, Christos Kozyrakis, HotOS ‘21: Proceedings of the Workshop on Hot Topics in Operating Systems, “A case against (most) context switches” - 2021-06-03

Significance of Process Management

Discuss who will benefit from exploring this topic and how they will benefit

Importance of Process Management:

  1. Better overall system performance and responsiveness
  2. Efficient resource allocation
  3. Maintained system stability
  4. Energy efficiency and sustainability

Beneficiaries:

  1. Users - better experience as a result of enhanced system performance
  2. Businesses - increasing efficiency and sustainability demands in digital services

To identify (some optional):

  1. Identify problem found
  2. Key concepts
  3. Identify solution developed and results
  4. Strength and weakness
  5. Identify gaps in research

Kernel Context Isolation and Context Switching

  1. Problem
    1. It is important to isolate application components for security
      1. they safeguard sensitive data and code
    2. There are massive performance costs to isolating application components using the process-based approach—the commonly used approach to application isolation
    3. The current solutions are insufficient
      1. stuff like kernel abstraction and CPU extensions, and etc. does not isolate enough kernel resources like file descriptors
  2. Solution
    1. They introduce a kernel isolation mechanism that includes memory protection
    2. This sufficiently isolates kernel resources with much lesser performance overhead
  3. Results and assessment
    1. After testing it on Firefox libraries and HTTP servers, they claim that the isolation overhead got reduced by about 32.7-98.4% when compared to other techniques
    2. They first tried it on the media decoding and Gzip decompression features of Firefox
    3. They then observed the latency and resource utilization of the technique on loading pages using Firefox
    4. They measured HTTP Server performance using a benchmarking tool called bombardierto. They discover that their solution results in significantly less overhead.
    5. From a security standpoint
      1. the usage of Intel MPK provides memory isolation and guarantees hardware memory protection
      2. Their implementation demonstrates that context switching can be safely done when using the Write Data to User Page Key Register (WRPKRU) and Restore Processor Extended States (XRSTOR) instructions.
      3. They were able to show that attackers cannot exploit several system calls to escape their kernel context isolation.
  4. Gaps
    1. MPK combined with their solution is limited by the number of concurrent muContext’s per address space.
    2. It requires refactoring effort from developers.
    3. It can still be extended to support workloads where IPC overhead is important

Why Context Switching Should Mostly Be Eliminated

  1. Problem
    1. Frequent context switches causes high overheads and severely worsens the complexity of kernels and applications
    2. Regular execution stream is disrupted as a result and makes it costly to perform save and resume states
  2. Solution
    1. Hardware threading model should be radically changed to focus on software problems instead of hardware ones
    2. It makes thread multiplexing obsolete by introducing a lot of hardware threads to each physical core
    3. ISA extensions is also presented to let the kernel and user software to leverage the threading model:
      1. this allows developers to eliminate interrupt and implement fast I/O without poling
      2. leads to simpler distributed programming models
      3. fast hypervisors
      4. system without exceptions
  3. Benefits
    1. The work involved with interrupt processing is avoided. An alternative means would let kernel designate hardware thread per core per interrupt type.
    2. Polling becomes unnecessary because this model would reduce the resource penalties of notifying waiting threads. A lot of threads per core would make polling useless because polling would only complicate core allocation.
    3. The threading models allows for system calls and VM-exits to operate in dedicated hardware threads without the mode switching overheads.
    4. Their model allows for kernel code to execute in one hardware thread and application code in a different hardware thread, thereby also allowing for kernels to use FP and vector operations and link with libraries, while also not having an effect on system call invocation latency
    5. Microkernel and container proxies performance and speed is improved by mitigating the excessive scheduling delays through a lot of software-managed threads.
    6. This model can allow isolation and alleviate code restrictions without sacrificing performance, unlike Hypervisors on today’s CPUs that rely on privileged access to host OS in order to have good performance
    7. The lack of hardware threads per physical cores led to distributed applications to depend on multiplex software threads on top of hardware threads. However, with much more hardware threads, distributed programming can be simplified, wherein developers can assign one hardware thread per request while avoiding thread scheduling overheads.

Android IPC

  1. Problem
    1. Android’s IPC mechanism implementation is vulnerable to intent redirection
    2. Malicious applications can start Android intent redirection and intent hijacking attacks to threaten user data confidentiality and integrity.
    3. Intent redirection involves malicious applications exploiting the publicly available exported components of legitimate applications to send malicious data to their components and services.
    4. These components and services receiving the malicious data may be involved with sensitive data operations, and, as such, exploiting them can allow hijacking attacks to gain unauthorized access to private data.
    5. the current Android IPC mechanism does not validate the identity of the IPC source and/or destination entities
  2. Solution
    1. IntentAuth was developed to provide an IPC mechanism with better security and a feature that allows users to control application interactions (user-defined policies)
    2. It provides encryption for intent data transmission
    3. It provides a way to authenticate the source/destination of inter-process communication
    4. It supports secure key storage components
  3. Benefits
    1. Because the weakest link of security for Android applications is the developers, this mechanism takes a user-centric approach independent of static policies defined by developers. It gives additional security services to strengthen the security of inter-process communication and enforce authentication and encryption.

Middleware Technologies and IPC

  1. Significance
    1. Distributed systems have proliferated and are gradually becoming more embedded in our everyday lives. They are involved in data processing, medical science, industrial control systems, aerospace, and enterprise applications.
    2. IPC is a fundamental component of distributed systems.
    3. This paper aims to evaluate the performance of several IPC methods
  2. Findings
    1. The socket as basic IPC yielded the best results in terms of speed
      1. The reason being is they avoid applying additional operations on the input data that will be sent.
    2. Nonetheless, their performance is worsened as a result of adding additional layers
    3. Additional layers are necessary for extra features, such as the elimination of platform dependencies and asynchronous communication

Graph View