Increasing emphasis on better networking in the Cloud
There has been a lot of discussion about how cloud computing is
changing the networking paradigm from the traditional LAN/WAN end-to-end
service guarantee to the public cloud “best effort” service delivery.
This can for example be viewed in the success of services like
Salesforce.com.
Salesforce.com has ca. 3 million customers that use their solutions to support critical sales and support processes. Regardless of the importance of the services provided, salesforce.com does not provide customers any availability guarantee and, since services are typically delivered over the Internet, no end-to-end guarantees can be provided either.
Is best effort good enough?
The “best effort” level of obtaining services from cloud service providers appears to becoming increasingly acceptable for a large share of companies. Especially is this true when it comes to SaaS services like salesforce.com and Google Apps for Business. These services seem to provide “good enough” service levels and not contradict with the number one concern of Cloud service adoption, namely, the concern about the security and confidentiality of data.
When it comes to public cloud services like IaaS the picture might become somewhat different however. Although many cloud users accept the inherent limitations of the Internet, including latency and packet-loss, when obtaining common IaaS services like compute and storage, increasingly more companies are requiring better networking alternatives from the perspective of reliability, security and speed.
Considering dedicated network connections to the Cloud
AWS, for example, have for a while offered the “Direct Connect” service which enables users to establish a dedicated network connection between their office network or on-premise data center, to one of AWS Direct Connect data center locations. This, essentially, means to establish a multi-partitioning VLAN via dedicated connection obtained through any of AWS partner network service providers. This is of course a step beyond establishing a VPN connection as with dedicated networking a certain level of performance, consistency and reliability can be assumed. Furthermore, this scheme also allows users to extend their private IP space into the Cloud while maintaining a network separation between public and private environments. For this purpose, AWS has partnered with several Tier-1 network providers in establishing global connection points for customers wishing to leverage the Cloud in this manner. Customers pay for the “per-hour” usage of selected bandwidth in terms of “port size”, i.e. 1Gb or 10Gb port at the nearest or appropriate connection point and the amount of data transfer in GB.
Although the networking issue has not been one of the primary inhibitors of IaaS adoption, it clearly plays an important role. Other, still more fundamental, factors have topped the list of concerns, including concerns about security and confidentiality of company data. When these have been addressed sufficiently, I would not be surprised to see the networking issue becoming more prominent.
Salesforce.com has ca. 3 million customers that use their solutions to support critical sales and support processes. Regardless of the importance of the services provided, salesforce.com does not provide customers any availability guarantee and, since services are typically delivered over the Internet, no end-to-end guarantees can be provided either.
Is best effort good enough?
The “best effort” level of obtaining services from cloud service providers appears to becoming increasingly acceptable for a large share of companies. Especially is this true when it comes to SaaS services like salesforce.com and Google Apps for Business. These services seem to provide “good enough” service levels and not contradict with the number one concern of Cloud service adoption, namely, the concern about the security and confidentiality of data.
When it comes to public cloud services like IaaS the picture might become somewhat different however. Although many cloud users accept the inherent limitations of the Internet, including latency and packet-loss, when obtaining common IaaS services like compute and storage, increasingly more companies are requiring better networking alternatives from the perspective of reliability, security and speed.
Considering dedicated network connections to the Cloud
AWS, for example, have for a while offered the “Direct Connect” service which enables users to establish a dedicated network connection between their office network or on-premise data center, to one of AWS Direct Connect data center locations. This, essentially, means to establish a multi-partitioning VLAN via dedicated connection obtained through any of AWS partner network service providers. This is of course a step beyond establishing a VPN connection as with dedicated networking a certain level of performance, consistency and reliability can be assumed. Furthermore, this scheme also allows users to extend their private IP space into the Cloud while maintaining a network separation between public and private environments. For this purpose, AWS has partnered with several Tier-1 network providers in establishing global connection points for customers wishing to leverage the Cloud in this manner. Customers pay for the “per-hour” usage of selected bandwidth in terms of “port size”, i.e. 1Gb or 10Gb port at the nearest or appropriate connection point and the amount of data transfer in GB.
Although the networking issue has not been one of the primary inhibitors of IaaS adoption, it clearly plays an important role. Other, still more fundamental, factors have topped the list of concerns, including concerns about security and confidentiality of company data. When these have been addressed sufficiently, I would not be surprised to see the networking issue becoming more prominent.
Categories:
CLOUD COMPUTING
0 comments:
Post a Comment