Precarious, Informalizing, and Flexible Work Transforming Concepts and Understandings
Summary (3 min read)
Introduction
- Multi-tier Heterogeneous Network where massive number of small cells are deployed under the coverage of macro cell(s) have been identified as one of the major approaches to enhance the capacity of 5G networks [1].
- The temporal and spatial variation of traffic load in cellular networks due to varying user distribution and mobility patterns results in most BSs been either idle (i.e. serving no user) or under-utilized (i.e. serving very few users) for most part of the day which amounts to energy wastage.
- This is not the case in the HetNet with Control/Data Separated Architecture (CDSA) where the macro BS provides continuous coverage, signalling functions and low data rate services while the small cells handles high data rate transmissions.
- Therefore, the authors propose a reinforcement learning based cell switching scheme that employs vertical traffic offloading to turn off lightly loaded small cell(s) in a HetNet with CDSA during periods of low traffic load.
- The rest of the paper is organized as follows.
A. Network Model
- The authors consider a two-tier HetNet consisting of a macro cell and three small cells as shown in Fig. 1, with separated control plane and data plane where the macro cells and small cells operate in dedicated frequency bands.
- The macro cell is responsible for providing coverage, signalling as well as low data rate services while the small cells provide high capacity in hotspot locations and are linked to the macro cell using the backhaul.
- Small cells are switched off during low traffic load periods and their traffic offloaded to their associated macro cell provided there is enough capacity in the macro cell to accommodate the offloaded traffic load.
B. Power Consumption Model of HetNet
- The authors assume a network consisting of BSs (both macro and small cells) each having limited number of resource blocks (RB) and that both macro cell and small cell have the same number of RBs.
- The authors model the load profile of each BS as the proportion of RBs occupied per minute over a 24 hour period.
- The total power consumption of the HetNet is the sum of the power consumption of the macro cell(s) and all the smalls cells under its coverage.
IV. PROBLEM FORMULATION
- The aim is to determine the optimal strategy to switch off lightly loaded small cells during low traffic periods that will minimise the total power consumption of the HetNet while considering the QoS requirement of offloaded traffic which is availability of capacity.
- P ( ψ ) is the expected power consumption of the HetNet using any switching strategy ψ.
- The second constraint is the dynamic power (power consumption due to transmission) constraint which implies that the power consumption gain ∑.
- P ks-off obtained by switching off small BS(s) must be greater than the increase in power consumption in the macro cell, ∆Pm as a result of additional load from sleeping small BS(s).
- The authors develop a reinforcement learning based small cell switching and traffic offloading mechanism in the next session to optimize energy consumption in the HetNet.
V. PROPOSED METHODOLOGY
- The authors propose a reinforcement learning algorithm, a set of machine learning algorithms, to implement the small cell switching operation.
- As a illustration in their study, the macro cell interacts with the network environment, obtains information about the traffic loads levels of the small cells through its backhaul connection with them and then decide which combination of small cells to switch off per time.
- Q-learning is an off-policy method, meaning that it follows different policies in determining the next action and updating the action-value table.
- The state is related to their optimization constraint which is availability of capacity in the macro cell for traffic offloading.
VI. PERFORMANCE EVALUATION
- The authors evaluate the performance of the proposed Q-learning based cell switching algorithm.
- For each simulation, the authors specify the maximum load level of the macro BS while the load of the small cells are continually varied.
- This is because Q-learning is able select the optimal set of small cell(s) to be switched off per time thereby enabling the HetNet to operate with minimal energy consumption.
- Fig. 3 presents the gain in energy consumption as well as the confidence levels of the results obtained when Q-learning is implemented.
- Initializing the action-value table at every episode could be an approach for this kind of abruptly changing environments; however, it comes with the expense of computational burden, since it makes Q-learning try to learn the environment at each episode.
VII. CONCLUSION
- The authors developed a Q-learning algorithm to minimize the total power consumption of a two-tier HetNet with CDSA.
- The Q-learning based cell switching algorithm is able to select the optimal set of small cells to be switched off in order to maximize energy saving in the HetNet.
- The increase in power consumption in the macro cell due to offloaded traffic from sleeping smalls cells was also considered when developing the switching mechanism.
- The result of the simulation reveals an energy savings gain of about 50% while ensuring that the QoS of users is maintained.
Did you find this useful? Give us your feedback
Citations
[...]
2,001 citations
1,773 citations
335 citations
References
10,062 citations
8,455 citations
"Precarious, Informalizing, and Flex..." refers background in this paper
...On the other, capital’s increased global mobility and need for flexibility is a reaction to the wave of struggles and strength of labor in the industrialized countries in the 1960s and 1970s (Castells & Portes, 1989; Hardt & Negri, 2001; Harvey, 2005; Silver, 2003)....
[...]
...As discussed in the introduction, a result of the neoliberal offensive on labor is that the barriers to the free movement of capital were removed (Harvey, 2005)....
[...]
[...]
3,822 citations
2,473 citations
[...]
2,282 citations
Related Papers (5)
Frequently Asked Questions (14)
Q2. Why do the salariat feel detached from the state social protection system?
Because of high incomes and identification with management, employers, the elite, and proficians, members of the salariat feel detached from the state social protection system, seeing their future security in terms of private insurance benefits, and earnings from judicious investment.
Q3. Why is precarious work being studied in the academy?
In the process, precarious work, as opposed to a more general social and existential precarity, is being studied in ways similar to flexible and casual labor.
Q4. What is the main contribution of Standing to the study of precarious work?
4 Standing’s (2008, 2011b) primary contribution to studying precarious work is the introduction of seven categories that serve as an alternative to the traditional working class-capitalist divide that is ill-equipped to accurately represent the complex layers of social-labor formations in contemporary global capitalism.
Q5. What is the meaning of the term precarious work?
The growing power and reach of global capital has exceeded the ability of nations and labor movements to regulate it, exacerbating inequality and precarious work.
Q6. What is the impact of precarious work on workers?
Policy makers’ and development organizations’ preoccupation with growth has numerous impacts for workers, in particular as they relate to the perceived need for greater flexibility (Kalleberg & Hewison, 2013).
Q7. What is the meaning of the term “precarious work”?
Standing’s typology pushes the conceptualization of precarious work beyond relation to social security, pension benefits, and other state-backed social protections.
Q8. What does the author try to do to minimize the uncertainty for the employer?
What individualizedat Maastricht University on April 7, 2013abs.sagepub.comDownloaded from298 American Behavioral Scientist 57(3)contracts often attempt to do is to tighten the conditions to minimize the uncertainty for the employer and to maximize the capacity to impose penalties for abrogation of the terms of the labor agreement.
Q9. What is the result of the neoliberal offensive on labor?
As discussed in the introduction, a result of the neoliberal offensive on labor is that the barriers to the free movement of capital were removed (Harvey, 2005).
Q10. What is the need to understand the precarious labor movement?
there is a need to better understand the ambitions, desires and strategies of precarious workers’ efforts in organizing and the broader implications of social struggle for alternatives to the dominant development paradigms.
Q11. Why is precarity used in the academic literature?
Precarity has most prominently been used as an accusation against the reduction or elimination of certain, primarily state backed social protections common among formerly Keynesian states (Arnold & Aung, 2011).
Q12. What are the main reasons for the lack of public and private investment in skills and development?
The lack of public and private investment in skills and development is accompanied by a lack of access to schooling, where women and ethnic and racial minorities disproportionally bear the brunt of these disadvantages (ILO, 2005; Kalleberg, 2009, 2011; Standing, 2008; Vosko, MacDonald, & Campbell, 2009).
Q13. What was the prominent argument for the informal economy?
It was particularly prominent when support for Western-led approaches to economic development, or the “Washington Consensus,” was strong.
Q14. What is the result of the unifying view?
A consequence of this unifying view is that the debate among certain scholars and institutions has shifted towards an assessment of the relative size of the different segments and the factors that influence them (Bacchetta et al., 2009).