Configuration management at massive scale: system design and experience

Enck, William; Moyer, Thomas; McDaniel, Patrick; Sen, S.; Sebos, P.; Spoerel, Sylke; Greenberg, Albert; Sung, Yu-Wei Eric; Rao, Sanjay; Aiello, William

doi:10.1109/jsac.2009.090408

Cited by 57 publications

(58 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…End-to-end Class of Service (CoS) is provisioned by marking packets and treating them differently according to their markings, on the CER-PER-backbone-PER-CER path. The dynamic and heterogeneous nature of changes to CERs [11] makes them the focus of our study.…”

Section: Algorithm For Extracting Correlated Changesmentioning

confidence: 99%

See 1 more Smart Citation

Extracting Network-Wide Correlated Changes from Longitudinal Configuration Data

Sung

Rao

Sen

et al. 2009

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. IP network operators face the challenge of making and managing router configuration changes to serve rapidly evolving user and organizational needs. Changes are expressed in low-level languages, and often impact multiple parts of a configuration file and multiple routers. These dependencies make configuration changes difficult for operators to reason about, detect problems in, and troubleshoot. In this paper, we present a methodology to extract networkwide correlations of changes. From longitudinal snapshots of low-level router configuration data, our methodology identifies syntactic configuration blocks that changed, applies data mining techniques to extract correlated changes, and highlights changes of interest via operator feedback. Employing our methodology, we analyze an 11-month archive of router configuration data from 5 different large-scale enterprise Virtual Private Networks (VPNs). Our study shows that our techniques effectively extract correlated configuration changes, within and across individual routers, and shed light on the prevalence and causes of systemwide and intertwined change operations. A deeper understanding of correlated changes has potential applications in the design of an auditing system that can help operators proactively detect errors during change management. To demonstrate this, we conduct an initial study analyzing the prevalence and causes of anomalies in system-wide changes.

show abstract

Section: Algorithm For Extracting Correlated Changesmentioning

confidence: 99%

“…However, they focus on detecting misconfiguration using static configuration snapshots. Several works have sought to automate top-down generation of low-level configurations [13,11] and typically focus on greenfield deployments. Others [8,9,3] have looked at detailed modeling and detection of errors in static configuration snapshots.…”

Section: Related Workmentioning

confidence: 99%

Extracting Network-Wide Correlated Changes from Longitudinal Configuration Data

Sung

Rao

Sen

et al. 2009

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…A few works have developed configuration templates in different networking fields (e.g., ISP [4,6,7]), and the present work shares the spirit of the use of templates. In contrast to those work, an originality of this work is the focus on the fundamental and crucial difficulty in creating appropriate templates.…”

Section: Related Workmentioning

confidence: 99%

“…Tenant construction in such datacenters still mostly relies on manual configuration because of the lack of tools that generally meet those various requirements. Manual configuration is substantially time-consuming and error-prone in general as claimed over at least a decade [3,4,5,6,7], and the use of multi-tenancy in datacenter context has unfortunately required more careful and sophisticated management processes than ever due to the increase in risks of misconfigurations in such shared environments.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Discovering configuration templates of virtualized tenant networks in multi-tenancy datacenters via graph-mining

Himura

Yasuda

2012

SIGCOMM Comput. Commun. Rev.

View full text Add to dashboard Cite

Multi-tenant datacenter networking, with which multiple customer (tenant) networks are virtualized over a single shared physical infrastructure, is cost-effective but poses significant costs on manual configuration. Such tasks would be alleviated with configuration templates, whereas a crucial difficulty stems from creating appropriate (i.e., reusable) ones. In this work, we propose a graph-based method of mining configurations of existing tenants to extract their recurrent patterns that would be used as reusable templates for upcoming tenants. The effectiveness of the proposed method is demonstrated with actual configuration files obtained from a business datacenter network.

show abstract

Towards centralized and semi‐automatic VLAN management

Yang

et al. 2014

Int J Network Mgmt

View full text Add to dashboard Cite

Virtual local area networks (VLANs) are widely used in many enterprises, campuses, and data-centre networks. Although VLANs can restrict broadcast domains and contain hosts in one network or separate networks, the management of VLANs is an ad hoc and error-prone work. In this paper, we design and implement a centralized and semi-automatic system for VLAN management (CSS-VM). Based on the physical network topology and user group (examples are engineering, student cluster, and faculty cluster) information, CSS-VM can decide the number of VLANs that each user group would be partitioned into. In addition, CSS-VM can not only monitor the status of devices and links but can also calculate an optimal spanning tree for each affected VLAN after a link or device failure is detected, so that it does not need to enable the Spanning Tree Protocol on devices but still has the ability of converging from device or link failures and avoiding bridging loops. We have evaluated CSS-VM on the topology and VLAN partition data of an operational enterprise network. Our results show that CSS-VM can obviously keep the broadcast traffic cost reasonable, efficiently partition and configure VLANs, quickly converge from link and device failures, and intelligently make a balanced use of links.Virtual local area networks (VLANs) play an important role in enterprises and campus networks. Enterprise or campus network operators often group the users based on the roles they are playing. Users in a user group are in the same broadcast domain and have the same access permission, which will simplify the administration and security tasks. Theoretically, we can create one VLAN for a whole user group. However, if one user group has too many users, the broadcast domain of this VLAN will be relatively large. In this case, a mass of broadcast packets will occupy the bandwidth and then degrade network performance. In order to control the broadcast domain in a reasonable scope, further partition of such a user group is needed. Obviously, the more VLANs into which a user group is partitioned, the smaller the average broadcast domain of the VLANs is. However, limited by the hardware capacity, the total number of VLANs has an upper bound. Thus, it is important for the operators to make a trade-off between the hardware capacity and network performance.Currently, both the ad hoc fashion in VLAN design and the complicated configuration in VLAN implementation may cause poor network performance. In the design phase, operators choose strategies to partition VLANs in an ad hoc fashion. Without a holistic viewpoint, the division plan may lead to much broadcast traffic, underutilization of some high-performance links, and overload of other links, which often deviates from the optimal design. In the implementation phase, operators enable the parameters by manually inputting many commands. For a network that may contain many :52-73 TOWARDS CENTRALIZED AND SEMI-AUTOMATIC VLAN MANAGEMENT 53 devices produced by different manufacturers, operators must configure the trunk links...

show abstract

Configuration management at massive scale: system design and experience

Cited by 57 publications

References 13 publications

Extracting Network-Wide Correlated Changes from Longitudinal Configuration Data

Extracting Network-Wide Correlated Changes from Longitudinal Configuration Data

Discovering configuration templates of virtualized tenant networks in multi-tenancy datacenters via graph-mining

Towards centralized and semi‐automatic VLAN management

Contact Info

Product

Resources

About