Identification of the deliverable or milestone | |
---|---|
Project: EMI | Deliverable or milestone identifier: DNA1.2.1 |
Title: DNA1.2.1 - Service Level Agreement Template | Doc. identifier: EMI-DNA1.2.1-1277517-SLA_Template-v0.5.doc |
Author(s): Alberto Di Meglio | Due date: PM 2 |
Identification of the reviewer | ||
---|---|---|
Name: Francesco Giacomini | Affiliation: INFN | EMI Activity/External project or Institute: SA1 |
Review date | 10/10/2010 |
Author(s) revision date | 17/11/2010 |
Reviewer acceptance date | 29/11/2010 |
N° | Page | Section | Observations | Is Addressed? |
1 | 6 | 1.3 | [R3] I hope DSA1.1 will be released at the same time as this deliverable. -- FrancescoGiacomini - 2010-10-10 This I cannot really control, but I agree that it would be best to have DSA1.1 out first -- AlbertoDiMeglio - 2010-11-17 |
![]() |
2 | 6 | 1.4 | A paragraph on the amendment procedure of the SLA template would be useful. -- FrancescoGiacomini - 2010-10-10 Added -- AlbertoDiMeglio - 2010-11-17 |
![]() |
3 | 8 | 2 | I would remove all the footnotes, all the terms, including the URLs, are mentioned in the Terminology table. -- FrancescoGiacomini - 2010-10-10 It seems to me easier to look at the foot of the page rather than going back to the Terminology table at the front of the document. Since both reviewers have expressed this comment, I guess that resistance is futile and I've complied -- AlbertoDiMeglio - 2010-11-17 |
![]() |
4 | 10 | 3.1 | Footnote 7 may become a reference. -- FrancescoGiacomini - 2010-10-10 Done -- AlbertoDiMeglio - 2010-11-17 |
![]() |
5 | 10 | 3.2 | Is it really needed to describe the EMI Work Packages? -- FrancescoGiacomini - 2010-10-10 Yes, because they play different roles in the monitoring or application of the SLAs, I'd like these roles to be clear. I agree that the definitions could be added by reference, but in my opinion it is easier to have the text included rather than asking to look in another document -- AlbertoDiMeglio - 2010-11-17 |
![]() |
6 | 12 | 3.4 | IGE should appear in the Terminology table. -- FrancescoGiacomini - 2010-10-10 I guess you mean IGE. Added -- AlbertoDiMeglio - 2010-11-17 Yes, I really mean IGE. -- FrancescoGiacomini - 29-Nov-2010 |
![]() |
7 | 15 | 5 | September 2010 is already in the past. -- FrancescoGiacomini - 2010-10-10 Well, yes, that is the date of the submission to the internal reviewers, the date will have to be updated when this is released officially -- AlbertoDiMeglio - 2010-11-17 |
![]() |
8 | 21 | 4(SLA) | Are we really affected by fiscal years? -- FrancescoGiacomini - 2010-10-10 Removed "fiscal", just once per year -- AlbertoDiMeglio - 2010-11-17 |
![]() |
9 | 21 | 4(SLA) | "terminate the relationship if all parties cannot abide by the revisions". Shouldn't the previous revision still hold, unless expired? -- FrancescoGiacomini - 2010-10-10 Usually the revision expires and replaces the previous versions. A revision is made to fix some type of issue on either side, leaving the previous version in effect may not be possible or desirable. I've explicitly mentioned that new versions expire the previous ones -- AlbertoDiMeglio - 2010-11-17 |
![]() |
10 | 22 | 5.1(SLA) | I would expect that the bullets in the list would become paragraph/subsections in the following pages but I can't see them, apart from something on Web-based support. There should be at least a description of what each of those services means. Then the Customer responsibilities and the EMI Responsibilities should be structured around those services. -- FrancescoGiacomini - 2010-10-10 One of the most critical comments from the CERN legal services around the first version of the SLA template has been that the SLA document itself should not describe the services, just the associated service levels. The description of the services must be provided to the Customers in a separate document, which in our case is the EMI User Support Plan. This is supported also by my analysis of a number of SLA from companies providing services to CERN. The Service Description and the SLA are always provided as separate documents. -- AlbertoDiMeglio - 2010-11-17 |
![]() |
11 | 22 | 5.2.1(SLA) | "Component" has not been defined. And in any case I would replace it with "Product". -- FrancescoGiacomini - 2010-10-10 Added definition of Component in the Glossary. However, I'm not sure Component and Product are equivalent. A Product is a generic EMI Middleware Service, Client or Library, but it cam be composed of several Components -- AlbertoDiMeglio - 2010-11-17 |
![]() |
12 | 22 | 5.2.1(SLA) | Using "you" to mean the customer is inappropriate, unless it is defined before, e.g. in Section 3 (of the SLA) in correspondence with Customer. -- FrancescoGiacomini - 2010-10-10 Indeed, it should be Customer everywhere. I replaced all the occurences of "you" were appropriate -- AlbertoDiMeglio - 2010-11-17 |
![]() |
13 | 22 | 5.2.2(SLA) | Where are (or will be) the Designated Contacts specified? Maybe in Section 11 (of the SLA), "Communication"? If so, tell it. -- FrancescoGiacomini - 2010-10-10 Yes, it is section 11. I've explicitly added the word "designated" -- AlbertoDiMeglio - 2010-11-17 |
![]() |
14 | 23 | 5.2.5(SLA) | Bullet (a): is it reasonable to require to backup and secure all programs and data on affected services? -- FrancescoGiacomini - 2010-10-10 I would say so, we need to protect us. If we suggest a workaround or a fix and this causes loss of data, we cannot be held responsible. This is a standard clause everywhere. If you check the instructions of most software updates, service packs, etc, there is always written something to that effect. We could explicitly say that whatever happens is not our fault, but having a concrete action as a reference is safer. if someone says: "You completly lost three years of analysis data!!!", I believe it is better to ask "Did you run the backup before installing the fix?" than saying "Sorry, not my fault. Thanks for using EMI" -- AlbertoDiMeglio - 2010-11-17 |
![]() |
15 | 23 | 5.2.5(SLA) | Bullet (b): it is not about "Backup and Removal of Data", which is the title of the section. -- FrancescoGiacomini - 2010-10-10 Not sure I understand this. It talks about backup and removal of data and software, I've added "and Software" to be more specific, but the rest looks fine to me -- AlbertoDiMeglio - 2010-11-17 |
![]() |
16 | 23 | 5.3(SLA) | The link http://www.eu-emi.eu/support/EMI_Support_Plan.pdf![]() -- FrancescoGiacomini - 2010-10-10 True, I need the official version of the Support Plan before i can upload it to the web. Hopefully it will be there before this is submitted -- AlbertoDiMeglio - 2010-11-17 |
![]() |
17 | 24 | 5.3.2(SLA) | The link http://www.eu-emi.eu/services/lifecycle/support_periods.htm![]() -- FrancescoGiacomini - 2010-10-10 Same as above, we need to define those support periods for each Product. i have to mention links to this artefacts, but I cannot provide the content. We can put an overall EMI Releases support policy statement, as soon as the policy is defined in the EMI Support Plan -- AlbertoDiMeglio - 2010-11-17 |
![]() |
18 | 25 | 5.3.5(SLA) | Can EMI charge any cost to a Customer? -- FrancescoGiacomini - 2010-10-10 Yes. For example, in case we are asked to have someone provide on-site technical support, we can ask for travel and subsistence to be paid by the Customer. We have no special provision for this in the EMI budget. The travel and subsistence budget is usually for attending EMI events or conference, workshops, etc. -- AlbertoDiMeglio - 2010-11-17 |
![]() |
19 | 25 | 5.4(SLA) | The link http://www.eu-emi.eu/support/EMI_Support_Plan.pdf![]() -- FrancescoGiacomini - 2010-10-10 See comment 16 -- AlbertoDiMeglio - 2010-11-17 |
![]() |
20 | 27 | 5.4.1(SLA) | Will we really provide Product Forums? it seems to me a lot of work and overlaps with already established communication channels for the communities around the existing services. -- FrancescoGiacomini - 2010-10-10 This is something that was foreseen by NA2. The liferay system INFN is setting up has functionality to host forums of this kind. This are sort of users-help-users forums, not standard support channels. Most companies have this self-help tools in place, it helps building an expert community and decreases the load on the official support channels -- AlbertoDiMeglio - 2010-11-17 |
![]() |
21 | 27 | 4.3(SLA) - but should be 5.4.2 | Is this section about the normal User Support or is it something else? -- FrancescoGiacomini - 2010-10-10 Updated section number. The section is supposed to be about the standard user support. I can modify it, if you give me some more specific comment -- AlbertoDiMeglio - 2010-11-17 |
![]() |
22 | 27 | 6.1(SLA) | The text of the Comments for the "Web Support" Service needs some rewriting. -- FrancescoGiacomini - 2010-10-10 I would need some more specific comment on this. This is the current writing "Web support requests are automatically acknowledged upon reception. Requests received after office hours will be stored in the support system, however no action can be guaranteed until the next working day" -- AlbertoDiMeglio - 2010-11-17 |
![]() |
23 | 28 | 6.2(SLA) | GGUS doesn't have the concept of Severity, only of Priority, which is a field set by the user during submission, with 4 possible values. There is no criteria on how to select a certain value. -- FrancescoGiacomini - 2010-10-10 Actually, GGUS by definition of user support system has only Severity, the problem is that they call it Priority and give the Priority levels a mix of Priority and Urgency values. It's rather messy. I suggest we interpret the GGUS field as Severity and map the values to the more standard Severity values that we want to use. I've added a map of values in the table for clarity -- AlbertoDiMeglio - 2010-11-17 |
![]() |
24 | 29 | 9.1(SLA) | The probability of "acts of God" is so small that we could ignore them. -- FrancescoGiacomini - 2010-10-10 The term "acts of God" is an official legal term in Contract Law used as an "implied defense under the rule of impossibility: i.e, the promise is discharged because of unforeseen, naturally occurring events that were unavoidable and which would result in insurmountable delay, expense, or other material breach" -- AlbertoDiMeglio - 2010-11-17 |
![]() |
25 | 29 | 9.2(SLA) | What does "assignment" mean in this context? -- FrancescoGiacomini - 2010-10-10 In legal parlance an Assignment is "a transfer of rights between two parties". This clause means that the Customer cannot transfer its support rights to another party. For example if we have an SLA with a project for one year, when the project ends the SLA cannot be transferred to another projects carrying over the same activities without prior written consent of EMI -- AlbertoDiMeglio - 2010-11-17 |
![]() |
26 | 33 | Appendix: Definitions | The definition of Severity is slightly different than the one used in DSA1.1 (Maintenance and User Support Plan), merging the concepts of Severity and Impact as introduced in DSA1.1. This is not a big deal in fact; it's more important to address comment #23, on matching what GGUS offers. -- FrancescoGiacomini - 2010-10-10 I haven't found a detailed definition of the Severity levels in DSA1.1 (which BTW, I think it's something missing from the deliverable, there should be a more precise definition of the Severity levels in the User Support section). The only definition I've found is "Severity: a measure of the degradation of the quality of service of the affected component". The definition of the severity levels in the SLA Template doesn't seem to be in contradiction with this generic definition of Severity in DSA1.1 -- AlbertoDiMeglio - 2010-11-17 |
![]() |