diff options
author | Gregory Glover <gg2147@att.com> | 2017-11-13 13:52:58 +0000 |
---|---|---|
committer | Gerrit Code Review <gerrit@onap.org> | 2017-11-13 13:52:58 +0000 |
commit | b284df49a1eafe673397f8cb6a26e4313c8b4dd5 (patch) | |
tree | 2e6671fab900c85a1d4bd1e7f25f67b1e769b126 /docs | |
parent | 00076c6822b526dc3962941798113bec96ac5ff6 (diff) | |
parent | 644d806209b652dd037ff27498629bfa9f9ca63c (diff) |
Merge "Fix figure numbering"
Diffstat (limited to 'docs')
-rw-r--r-- | docs/guides/onap-developer/architecture/onap-architecture.rst | 26 |
1 files changed, 14 insertions, 12 deletions
diff --git a/docs/guides/onap-developer/architecture/onap-architecture.rst b/docs/guides/onap-developer/architecture/onap-architecture.rst index 903303fef..95b91b2d9 100644 --- a/docs/guides/onap-developer/architecture/onap-architecture.rst +++ b/docs/guides/onap-developer/architecture/onap-architecture.rst @@ -1,4 +1,4 @@ -.. This work is licensed under a Creative Commons Attribution 4.0 International License. +.. This work is licensed under a Creative Commons Attribution 4.0 International License. .. http://creativecommons.org/licenses/by/4.0 .. Copyright 2017 Huawei Technologies Co., Ltd. @@ -84,11 +84,11 @@ following foundational principles: services - Carrier-grade scalability including horizontal scaling (linear - scale-out) and distribution to support large number of services - and large networks + scale-out) and distribution to support large number of services + and large networks - Metadata-driven and policy-driven architecture to ensure flexible - ways in which capabilities are used and delivered + ways in which capabilities are used and delivered - The architecture shall enable sourcing best-in-class components @@ -97,7 +97,7 @@ following foundational principles: - Core capabilities shall support many diverse services - The architecture shall support elastic scaling as needs grow or - shrink + shrink |image0|\ @@ -115,7 +115,9 @@ necessary to develop service/operations-specific collection, analytics, and policies (including recipes for corrective/remedial action) using the ONAP Design Framework Portal. -|image1|\ **Figure 2:** ONAP Platform components (Amsterdam Release) +|image1|\ + +**Figure 2:** ONAP Platform components (Amsterdam Release) Portal ++++++ @@ -257,16 +259,16 @@ Closed-Loop Automation The following sections describe the ONAP frameworks designed to address these major requirements. The key pattern that these frameworks help -automate is +automate is: -***Design -> Create -> Collect -> Analyze -> Detect -> Publish -> -Respond.*** +**Design -> Create -> Collect -> Analyze -> Detect -> Publish -> +Respond** We refer to this automation pattern as “closed-loop automation” in that it provides the necessary automation to proactively respond to network and service conditions without human intervention. A high-level schematic of the “closed-loop automation” and the various phases within -the service lifecycle using the automation is depicted in Figure 4. +the service lifecycle using the automation is depicted in Figure 3. Closed-loop control is provided by Data Collection, Analytics and Events (DCAE) and other ONAP components. Collectively, they provide FCAPS @@ -322,7 +324,7 @@ providers to offer new value-added services to their customers with less dependency on the underlying hardware. In this use case, the customer has a physical CPE (pCPE) attached to a -traditional broadband network such as DSL (Figure 1). On top of this +traditional broadband network such as DSL (Figure 4). On top of this service, a tunnel is established to a data center hosting various VNFs. In addition, depending on the capabilities of the pCPE, some functions can be deployed on the customer site. @@ -368,7 +370,7 @@ To connect the different data centers, ONAP will also have to interface with legacy systems and physical function to establish VPN connectivity in a brown field deployment. -The VoLTE use case, shown in Figure 6, demonstrates the use of the VF-C +The VoLTE use case, shown in Figure 5, demonstrates the use of the VF-C component and TOSCA-based data models to manage the virtualization infrastructure. |