中文释义
网络演算(Networkcalculus)是一种基于非线性代数的确定性排队理论,目前已广泛应用于计算机网络建模与性能分析,特别是为计算延迟和积压等端到端性能参数的确界提供了有效工具。
网络演算的研究工作可分为理论研究与应用研究。理论研究的研究对象是网络演算的数学模型,是网络演算得以广泛应用的基础。理论研究的每一步进展都可以带动大批实际应用。由于网络演算属于跨学科的研究领域,涉及较多数学理论,因此理论研究的进展相对于应用而言较为缓慢。
网络演算作为一种网络性能分析工具,按照普遍的观点可以分为确定性网络演算和随机网络演算。确定性网络演算比较简单,目的是得到网络性能的最坏边界。随机网络演算的目的是为网络提供随机服务质量保障,要考虑网络数据流的随机突发,自相识等特性,以及网络信道的接入拥塞物理信道的衰落等因素,因此相对应用起来比较复杂,因此许多研究人员也运用了不同的数学方法和不同的数学表达方式来扩充随机网络演算。
网络演算的主要工具是到达曲线和服务曲线。
网络演算主要是用来得到以下5个结果:
1)数据流叠加
2)输入输出的关系
3)网络延时和积压性质
4)节点串联
5)剩余服务曲线
https://baike.baidu.com/item/%E7%BD%91%E7%BB%9C%E6%BC%94%E7%AE%97
http://gb.oversea.cnki.net/KCMS/detail/detail.aspx?filename=2010147286.nh&dbcode=CDFD&dbname=CDFDREF
英文释义
Network calculus is “a set of mathematical results which give insights into man-made systems such as concurrent programs, digital circuits and communication networks.”[1] Network calculus gives a theoretical framework for analysing performance guarantees in computer networks. As traffic flows through a network it is subject to constraints imposed by the system components, for example:
- link capacity
- traffic shapers (leaky buckets)
- congestion control
- background traffic
These constraints can be expressed and analysed with network calculus methods. Constraint curves can be combined using convolution under min-plus algebra. Network calculus can also be used to express traffic arrival and departure functions as well as service curves.
The calculus uses “alternate algebras … to transform complex non-linear network systems into analytically tractable linear systems.”[2]
Currently, there exists two branches in network calculus: one handling deterministic bounded, and one handling stochastic bounds .[3]
System modelling
Modelling flow and server
In network calculus, a flow is modelled as cumulative functions A, where A(t) represents the amount of data (number of bits for example) send by the flow in the interval [0,t). Such functions are non-negative and non-decreasing. The time domain is often the set of non negative reals.
Arrival and departure curve at ingress and egress (入口和出口) of a server.
A
:
R
+
→
R
+
A
:
R
+
→
R
+
{\displaystyle A:\mathbb {R} ^{+}\rightarrow \mathbb {R} ^{+}}A:{\mathbb R}^{+}\rightarrow {\mathbb R}^{+}
A:R+→R+A:R+→R+
∀ u , t ∈ R + : u < t ⟹ A ( u ) ≤ A ( t ) ∀ u , t ∈ R + : u < t ⟹ A ( u ) ≤ A ( t ) {\displaystyle \forall u,t\in \mathbb {R} ^{+}:u<t\implies A(u)\leq A(t)}\forall u,t\in {\mathbb R}^{+}:u<t\implies A(u)\leq A(t) ∀u,t∈R+:u<t⟹A(u)≤A(t)∀u,t∈R+:u<t⟹A(u)≤A(t)
A server can be a link, a scheduler, a traffic shaper, or a whole network. It is simply modelled as a relation between some arrival cumulative curve A and some departure cumulative curve D. It is required that A ≥ D, to model the fact that the departure of some data can not occur before its arrival.
Modelling backlog and delay
Given some arrival and departure curve A and D, the backlog at any instant t, denoted b(A,D,t) can be defined as the difference between A and D. The delay at t, d(A,D,t) is defined as the minimal amount of time such that the departure function reached the arrival function. When considering the whole flows, the supremum of these values is used.
Horizontal and vertical deviation between arrival and departure cumulative curves
b ( A , D , t ) : = A ( t ) − D ( t ) b ( A , D , t ) : = A ( t ) − D ( t ) {\displaystyle b(A,D,t):=A(t)-D(t)}b(A,D,t):=A(t)-D(t) b(A,D,t):=A(t)−D(t)b(A,D,t):=A(t)−D(t)
d ( A , D , t ) : = inf { d ∈ R + s . t . D ( t + d ) ≥ A ( t ) } d ( A , D , t ) : = inf { d ∈ R + s . t . D ( t + d ) ≥ A ( t ) } {\displaystyle d(A,D,t):=\inf \left\{d\in \mathbb {R} ^{+}~s.t.~D(t+d)\geq A(t)\right\}}d(A,D,t):=\inf \left\{d\in {\mathbb R}^{+}~s.t.~D(t+d)\geq A(t)\right\} d(A,D,t):=inf{d∈R+ s.t. D(t+d)≥A(t)}d(A,D,t):=inf{d∈R+ s.t. D(t+d)≥A(t)}
b ( A , D ) : = sup t ≥ 0 { A ( t ) − D ( t ) } b ( A , D ) : = sup t ≥ 0 { A ( t ) − D ( t ) } {\displaystyle b(A,D):=\sup _{t\geq 0}\left\{A(t)-D(t)\right\}}b(A,D):=\sup _{{t\geq 0}}\left\{A(t)-D(t)\right\} b(A,D):=t≥0sup{A(t)−D(t)}b(A,D):=supt≥0{A(t)−D(t)}
d ( A , D ) : = sup t ≥ 0 { inf { d ∈ R + s . t . D ( t + d ) ≥ A ( t ) } } d ( A , D ) : = sup t ≥ 0 { inf { d ∈ R + s . t . D ( t + d ) ≥ A ( t ) } } {\displaystyle d(A,D):=\sup _{t\geq 0}\left\{\inf \left\{d\in \mathbb {R} ^{+}~s.t.~D(t+d)\geq A(t)\right\}\right\}}d(A,D):=\sup _{{t\geq 0}}\left\{\inf \left\{d\in {\mathbb R}^{+}~s.t.~D(t+d)\geq A(t)\right\}\right\} d(A,D):=t≥0sup{inf{d∈R+ s.t. D(t+d)≥A(t)}}d(A,D):=supt≥0{inf{d∈R+ s.t. D(t+d)≥A(t)}}
In general, the flows are not exactly known, and only some constraints on flows and servers are known (like the maximal number of packet sent on some period, the maximal size of packets, the minimal link bandwidth). The aim of network calculus is to compute upper bounds on delay and backlog, based on these constraints. To do so, network calculus uses the min-plus algebra.
Min-plus algebra
In filter theory and linear systems theory the convolution of two functions f f f and g g g is defined as
( f ∗ g ) ( t ) : = ∫ 0 t f ( τ ) ⋅ g ( t − τ ) d τ {\displaystyle (f\ast g)(t):=\int _{0}^{t}f(\tau )\cdot g(t-\tau )d\tau } (f∗g)(t):=∫0tf(τ)⋅g(t−τ)dτ
In min-plus algebra the sum is replaced by the minimum respectively infimum operator and the product is replaced by the sum. So the min-plus convolution of two functions f f f and g g g becomes
( f ⊗ g ) ( t ) : = inf 0 ≤ τ ≤ t { f ( τ ) + g ( t − τ ) } {\displaystyle (f\otimes g)(t):=\inf _{0\leq \tau \leq t}\left\{f(\tau )+g(t-\tau )\right\}} (f⊗g)(t):=0≤τ≤tinf{f(τ)+g(t−τ)}
e.g. see the definition of service curves. Convolution and min-plus convolution share many algebraic properties. In particular both are commutative and associative.
A so-called min-plus de-convolution operation is defined as
( f ⊘ g ) ( t ) : = sup τ ≥ 0 { f ( t + τ ) − g ( τ ) } {\displaystyle (f\oslash g)(t):=\sup _{\tau \geq 0}\left\{f(t+\tau )-g(\tau )\right\}} (f⊘g)(t):=τ≥0sup{f(t+τ)−g(τ)}
e.g. as used in the definition of traffic envelopes.
The vertical and horizontal deviations can be expressed in terms of min-plus operators.
b ( f , g ) = ( f ⊘ g ) ( 0 ) {\displaystyle b(f,g)=(f\oslash g)(0)} b(f,g)=(f⊘g)(0)
d ( f , g ) = inf { w : ( f ⊘ g ) ( − w ) ≤ 0 } {\displaystyle d(f,g)=\inf\{w:(f\oslash g)(-w)\leq 0\}} d(f,g)=inf{w:(f⊘g)(−w)≤0}
Traffic envelopes
Cumulative curves are real behaviours, unknown at design time. What is known is some constraint. Network calculus uses the notion of traffic envelope, also known as arrival curves.
A cumulative function A is said to conform to an envelope (or arrival curve) E, if for all t it holds that
E ( t ) ≥ sup τ ≥ 0 { A ( t + τ ) − A ( τ ) } = ( A ⊘ A ) ( t ) {\displaystyle E(t)\geq \sup _{\tau \geq 0}\{A(t+\tau )-A(\tau )\}=(A\oslash A)(t)} E(t)≥τ≥0sup{A(t+τ)−A(τ)}=(A⊘A)(t).
Two equivalent definitions can be given
∀
τ
,
t
∈
R
+
:
A
(
τ
+
t
)
−
A
(
τ
)
≤
E
(
t
)
{\displaystyle \forall \tau ,t\in \mathbb {R} ^{+}:A(\tau +t)-A(\tau )\leq E(t)}
∀τ,t∈R+:A(τ+t)−A(τ)≤E(t) --------- (1)
A
≤
A
⊗
E
A\leq A\otimes E
A≤A⊗E --------- (2)
Thus, E places an upper constraint on flow A. Such function E can be seen as an envelope that specifies an upper bound on the number of bits of flow seen in any interval of length t starting at an arbitrary τ, cf. eq. (1).
Service curves
In order to provide performance guarantees to traffic flows it is necessary to specify some minimal performance of the server (depending on reservations in the network, or scheduling policy, etc.). Service curves provide a means of expressing resource availability. Several kinds of service curves exists, like weakly strict, variable capacity node, etc. See [4] [5] for an overview.
Minimal service
Let A be an arrival flow, arriving at the ingress of a server, and D be the flow departing at the egress. The system is said to provide a simple minimal service curve S to the pair (A,B), if for all t it holds that D ( t ) ≥ ( A ⊗ S ) ( t ) {\displaystyle D(t)\geq (A\otimes S)(t)} D(t)≥(A⊗S)(t).
Strict minimal service
Let A be an arrival flow, arriving at the ingress of a server, and D be the flow departing at the egress. A backlog period is an interval I such that, on any t ∈ I , A ( t ) > D ( t ) t ∈ I, A(t)>D(t) t∈I,A(t)>D(t).
The system is said to provide a strict minimal service curve S to the pair (A,B) iff, ∀ s , t ∈ R + {\displaystyle \forall s,t\in \mathbb {R} ^{+}} ∀s,t∈R+, such that s ≤ t {\displaystyle s\leq t} s≤t, if ( s , t ] {\displaystyle (s,t]} (s,t] is a backlog period, then D ( t ) − D ( s ) ≥ S ( t − s ) {\displaystyle D(t)-D(s)\geq S(t-s)} D(t)−D(s)≥S(t−s).
If a server offers a strict minimal service of curve S, it also offers a simple minimal service of curve S.
Basic results: Performance bounds and envelope propagation
From traffic envelope and service curves, some bounds on the delay and backlog, and an envelope on the departure flow can be computed.
Let A be an arrival flow, arriving at the ingress of a server, and D be the flow departing at the egress. If the flow as a traffic envelope E, and the server provides a minimal service of curve S, then the backlog and delay can be bounded:
b
(
A
,
D
)
≤
b
(
E
,
S
)
b(A,D)\leq b(E,S)
b(A,D)≤b(E,S)
d
(
A
,
D
)
≤
d
(
E
,
S
)
d(A,D)\leq d(E,S)
d(A,D)≤d(E,S)
Moreover, the departure curve has envelope E ′ = E ⊘ S . E'=E\oslash S. E′=E⊘S.
Moreover, these bounds are tight i.e. given some E, and S, one may build an arrival and departure such that b ( A , D ) = b ( E , S ) b(A,D) = b(E,S) b(A,D)=b(E,S) and v ( A , D ) = v ( E , S ) v(A,D)=v(E,S) v(A,D)=v(E,S).
Concatenation / PBOO
Consider a sequence of two servers, when the output of the first one is the input of the second one. This sequence can be seen as a new server, built as the concatenation of the two other ones.
Then, if the first (resp. second) server offers a simple minimal service S 1 S_{1} S1 (resp. S 2 S_{2} S2), then, the concatenation of both offers a simple minimal service S e 2 e = S 1 ⊗ S 2 S_{{e2e}}=S_{1}\otimes S_{2} Se2e=S1⊗S2.
Sequence of two servers
The proof does iterative application of the definition of service curves
X
≥
A
⊗
S
1
X\geq A\otimes S_{1}
X≥A⊗S1,
D
≥
X
⊗
S
2
D\geq X\otimes S_{2}
D≥X⊗S2 and some properties of convolution, isotonicity (
D
≥
(
X
⊗
S
2
)
⊗
S
1
D\geq (X\otimes S_{2})\otimes S_{1}
D≥(X⊗S2)⊗S1), and associativity (
D
≥
X
⊗
(
S
2
⊗
S
1
D\geq X\otimes (S_{2}\otimes S_{1}
D≥X⊗(S2⊗S1)).
The interest of this result is that the end-to-end delay bound is not greater than the sum of local delays: d ( E , S 2 ⊗ S 1 ) ≤ d ( E , S 1 ) + d ( E ⊘ S 1 , S 2 d(E,S_{2}\otimes S_{1})\leq d(E,S_{1})+d(E\oslash S_{1},S_{2} d(E,S2⊗S1)≤d(E,S1)+d(E⊘S1,S2).
This result is known as Pay burst only once (PBOO).
Tool
There are several tools based on network calculus.
The DiscoDNC is an academic Java implementation of the network calculus framework.[6]
The RTC Toolbox is an academic Java/MATLAB implementation of the Real-Time calculus framework, a theory quasi equivalent to network calculus.[4]
The CyNC[7] tool is an academic MATLAB/Symulink toolbox, based on top of the RTC Toolbox. The tool was developed in 2004-2008 and it is currently used for teaching at Aalborg university.
The RTaW-PEGASE is an industrial tool devoted to timing analysis tool of switched Ethernet network (AFDX, industrial and automotive Ethernet), based on network calculus.[8]
The Network calculus interpreter is an on-line (min,+) interpreter.
The WOPANets is an academic tool combining network calculus based analysis and optimization analysis.[9]
The DelayLyzer is an industrial tool designed to compute bounds for Profinet networks.[10]
DEBORAH is an academic tool devoted to FIFO networks.[11]
NetCalBounds is an academic tool devoted to blind & FIFO tandem networks.[12][13]
NCBounds is a network calculus tool in Python, published under BSD 3-Clause License. It considers rate-latency servers and token-bucket arrival curves. It handles any topology, including cyclic ones[14].
The Siemens Network Planner (SINETPLAN) uses network calculus (among other methods) to help the design of a PROFINET network.[15]
Reference
https://en.wikipedia.org/wiki/Network_calculus