Skip to content

Conversation

@marnikitta
Copy link
Contributor

No description provided.


\begin{definition}{Nullification time}
of an input element $a_\tau$ is the time $\theta_{a_\tau}=inf(\hat{\tau}>\tau|W_{\hat{\tau}}\setminus{S_{\hat{\tau}}}\cap{Cl(D)(a_\tau)=\emptyset})$, where $Cl(D)$ is a transitive closure of the relation $D$.
of an input element $a_\tau$ is the first time $\theta_{a_tau}$ when all elements that are produced from the $a_\tau$ are in state or are released from the system. $\theta_{a_\tau}=inf(\hat{\tau}>\tau|W_{\hat{\tau}}\setminus{S_{\hat{\tau}}}\cap{Cl(D)(a_\tau)=\emptyset})$, where $Cl(D)$ is a transitive closure of the relation $D$.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Убойная формула, расписал словами ее перед тем как формулу давать

Within our model, one can define streaming system using only data flow elements and business logic. However, without the notion of time, we cannot observe any processing progress. Let $\tau\in{\mathbb{N}}$ be an exact global discrete time. Let $a_\tau\in{\Gamma}$ be the element, which enters at the time $\tau$, and $b_\tau\in{2^\Gamma}$ be the elements, which leave at the time $\tau$. Let $A_{\tau}=\bigcup\limits_{i=1}^{\tau}{a_i}$ be a set of all input elements by the time $\tau$ and ${B}_\tau=\bigcup\limits_{i=1}^{\tau}{b_i}$ be a set of all output elements.
Basically $D$ captures the notion of a logical graph which vertices are pure functions.

Within our model, one can define streaming system using only data flow elements and business logic. However, without the notion of time, we cannot observe any processing progress. Let $\tau\in{\mathbb{N}}$ be an exact global discrete time. Let $a_\tau\in{\Gamma}$ be the element, which enters at the time $\tau$, and $b_\tau\in{2^\Gamma}$ be the elements, which leave at the time $\tau$. Let $A_{\tau}=\{a_i\}_{i=1}^{\tau}$ be a set of all input elements by the time $\tau$ and ${B}_\tau=\bigcup\limits_{i=1}^{\tau}{b_i}$ be a set of all output elements.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

a_i - элемент множества. Объединять элементы не кошено. Объединяют множества.

\end{definition}

Technically, states in existing stream processing systems are not data flow elements, but as it was mentioned above, they can be presented in such a way using drifting state model. The main purpose of the state is to accumulate the information about input items. Data flow elements cannot be in $W\setminus{S}$ for an infinite time by the definition. Hence, for each input element $a_\tau$, there is a nullification time $\theta_{a_\tau}$, thereafter all elements, which depend on $a_\tau$, are in the system state. Since the nullification time, the input element can affect output elements only through the state.
Technically, states in existing stream processing systems are not data flow elements, but as it was mentioned above, they can be presented in such a way using drifting state model. The main purpose of the state is to accumulate the information about input items. Data flow elements cannot stay in $W\setminus{S}$ for an infinite time by the definition. Hence, for each input element $a_\tau$, there is a nullification time $\theta_{a_\tau}$, thereafter all elements, which depend on $a_\tau$, are in the system state. Since the nullification time, the input element can affect output elements only through the state.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Что это за определение такое, что элементы не могут вечно сидеть в системе? По-хорошему, надо тогда вводить свойство прогресса в самом начале. А-ля "для любого a_i существует момент времени, когда он покинет систему"

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants