The goal of this Master’s thesis is to present the state of the art of the data centre network environment, together with the most recent solutions adopted and the current trends in the realization of data centre networks. The focus of the thesis is on the different Ethernet fabric solutions proposed in the recent years comparing the standard solutions (TRILL and SPB) from the standardization boards (IETF and IEEE), the proposals from the research field (PortLand, SEATTLE, VL2, improvements to TRILL), and the alternatives proposed by the vendors (FabricPath from Cisco, VLC form Brocade and QFabric from Juniper). Several experts from different companies were interviewed to gather insights and to better understand the current and future technologies adopted in data centres. From the interviews and the further comparison work emerges that, at the moment, several proprietary solutions are available and that companies willing to build a data centre easily face a vendor lock-in risk due to the several proprietary solutions available. This is why it is needed to carefully evaluate the needs of a company before deciding to adopt one of the solutions and understand if the benefits from a given proprietary solution are worth risking the lock-in. In general a company that already utilizes proprietary solutions from a single vendor might find easier to stick to that vendor, while typically it might be a better choice to stick to open standards until a unified solution is defined and published by the standardization boards.
L’obiettivo di questa tesi di laurea specialistica è quello di presentare lo stato dell’arte delle soluzioni adottate per la realizzazione di reti all’interno di un centro elaborazione dati. La tesi presenta anche un’analisi dettagliata di differenti soluzioni per la realizzazione di soluzioni Ethernet fabric proposte negli ultimi anni confrontando le soluzioni standard (TRILL e SPB) proposte dagli enti di standardizzazione (IETF e IEEE), le proposte dal mondo della ricerca (PortLand, SEATTLE, VL2, Miglioramenti per TRILL) e, da ultimo, le proposte di diverse aziende del settore (FabricPath per Cisco, VLC per Brocade e QFabric per Juniper). Parecchi esperti di diverse aziende sono stati intervistati per raccogliere informazioni utili per comprendere al meglio le tecnologie presenti e future utilizzate nei data centre. Dalle interviste e dal lavoro di confronto emerge come al momento siano disponibili parecchie tecnologie proprietarie e come aziende interessate alla realizzazione di un data centre possano facilmente essere costrette ad adottare soluzioni complete provenienti esclusivamente da un unico produttore (vendor lock-in). Ecco perché è necessario valutare attentamente le necessità di un’azienda prima di adottare uno di questi nuovi sistemi. In linea generale, un’azienda che già utilizza soluzioni proprietarie di un produttore può continuare a farlo ed adottare le soluzioni per Ethernet fabric proposte, altrimenti, se non vigono particolare requisiti da soddisfare, è meglio continuare ad implementare soluzioni standard in attesa della definitiva standardizzazione di una soluzione per reti Ethernet fabric.
State of the art of data centre network technologies case : comparison between Ethernet fabric solutions
MAESTRELLI, FRANCESCO
2010/2011
Abstract
The goal of this Master’s thesis is to present the state of the art of the data centre network environment, together with the most recent solutions adopted and the current trends in the realization of data centre networks. The focus of the thesis is on the different Ethernet fabric solutions proposed in the recent years comparing the standard solutions (TRILL and SPB) from the standardization boards (IETF and IEEE), the proposals from the research field (PortLand, SEATTLE, VL2, improvements to TRILL), and the alternatives proposed by the vendors (FabricPath from Cisco, VLC form Brocade and QFabric from Juniper). Several experts from different companies were interviewed to gather insights and to better understand the current and future technologies adopted in data centres. From the interviews and the further comparison work emerges that, at the moment, several proprietary solutions are available and that companies willing to build a data centre easily face a vendor lock-in risk due to the several proprietary solutions available. This is why it is needed to carefully evaluate the needs of a company before deciding to adopt one of the solutions and understand if the benefits from a given proprietary solution are worth risking the lock-in. In general a company that already utilizes proprietary solutions from a single vendor might find easier to stick to that vendor, while typically it might be a better choice to stick to open standards until a unified solution is defined and published by the standardization boards.| File | Dimensione | Formato | |
|---|---|---|---|
|
Thesis ITA.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Tesi
Dimensione
7.01 MB
Formato
Adobe PDF
|
7.01 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/43981