The only 2 points of interest that caught my attention in between a multitude of bullet points were that :
- 93% of companies want an interface between their ticketing system and the cloud because they want to be able to benchmark their remaining inhouse services against their cloud services. This is likely to be an important point for us as we move to BPOS, because while our ticket numbers will still be measured through GTS & Magic, our ability to track ticket resolution times etc. will be hindered by the fact that Magic will not integrate with whatever system Microsoft uses today, not in the short term, anyway. We need to make sure we have the required processes in place as BPOS is rolled out.
- 78% of companies want dedicated servers. Security is the main concern here, but mostly because the correlation (or lack of) between security and server isolation is not well understood by most companies. In an ideal world, yes we would have assessed multitenant vs. dedicated in detail, but given our timescales for making THE decision, it makes sense that we have gone with dedicated rather than multitenant.
Posts
3 comments:
I just noticed this post and was curious what you meant by the statements. I don’t know where to find the full presentation, but I got some comments based on the statements.
The first point seems to be a problem with insufficient tool integration between the cloud tooling and your internal process? You could always file the tickets through your normal process and then escalate them to the cloud provider. Since different providers have different cloud infrastructure, I’m sure there’ll be a market for integration points in the near term. And since clouds are pushing for automation, self-healing, self-servicing a lot of these requests are likely to be automatable or internally serviceable. Do you mind elaborating what more you are looking here?
The second point seems more controversial to me. Are you comparing on-premises against the cloud over the security of network control or between the cloud and traditional DC space renting? Each of them has different tradeoffs for security. In the first case, the control of the network is much stronger on-premises, but requires custom automation and maintenance to ensure continued security. With respect to renting dedicated space, the security of the machine seems to be more in the user’s space with all the negatives of shared infrastructure of the cloud. There are no automated updates of the OS or the network possible in the dedicated hosting since little is known of your requirements. A cloud provider has better knowledge of your stack and could provide continued network monitoring.
Note: I do work for MSFT, but cloud migration is something I wonder about.
Hi Aleks (quirkiest spelling I've seen of your name so far btw ;-) Thx for your comments.
Re. the first point - yes there is definitely a shortage of integration between enterprise ticketing systems and cloud vendors' ticketing systems at the moment. Without such integration, enterprises will still be able to track basic metrics like incident numbers etc. However, enterprises which are further along in their service management practices will be hindered by the lack of integration - eg. we track incident resolution time as one of our applications support performance metrics - this will neither be as easy to track nor as easily improvable once we move to the cloud, given that part of the resolution time will not be in our control.
About the second point, what I meant to say was that there is a perception in companies that cloud solutions are generally less secure than their own tried and tested on-prem solutions - but this perception is based more on the fact that there isn't yet any mainstream understanding of how on-prem security compares with cloud security. The presenter reckons that this point will be addressed as we have more and more trials of cloud computing, but in the meantime most companies, if they do venture out into the cloud, are much more likely to go for the dedicated model than for true multi-tenancy, which sort of hinders the cloud computing model given that the dedicated model is more of a hybrid of the external hosting model.
Very interested to know what it is you "wonder" about cloud migration! :)
(I think my previous response got lost)
I think the two points are heavily related because any type of outsourcing (dedicated hosting, private cloud, public cloud) will have the same problem with having external support. They would also have very similar security implication since they'd have shared infrastructure for cost reductions. I am not convinced that VM boundary is much different with respect to security.
The part that makes me wonder about cloud migration is statements like 'no more desktops in 3 years' (Google's european CXX I think), which imply a huge change in how clouds are used.
Post a Comment