First Bill for DLTJ Hosting on Amazon Web Services

I just got the bill for the first month of hosting this blog on Amazon Web Services. The total for the month was $23.60, and includes:

  • data transfer charges for all in-bound and out-bound content;
  • a full-time use of a LINUX micro-sized Elastic Compute Cloud (EC2) instance (with backup to the Elastic Block Store (EBS));
  • a Simple Storage Service (S3) bucket for static files (Cascading Style Sheet and JavaScript files, images, and other media); and
  • use of the Amazon CloudFront content distribution network.

All told, I’m pretty pleased with the costs — particularly as I was considering the amortized cost of buying a new server to replace the one I had been using for the past five years. The itemized bill is included below.

Billing Statement:February 1, 2011
The billing cycle for this report is is January 1 – January 31, 2011.

Amazon CloudFront

United States
  $0.150 per GB – first 10 TB / month data transfer out3.483 GB0.52
  $0.0100 per 10,000 HTTPS Requests4 Requests0.01
  $0.0075 per 10,000 HTTP Requests265,927 Requests0.20
  $0.150 per GB – first 10 TB / month data transfer out0.481 GB0.07
  $0.0090 per 10,000 HTTP Requests26,546 Requests0.02
  $0.201 per GB – first 10 TB / month data transfer out (includes consumption tax).0.054 GB0.01
  $0.0095 per 10,000 HTTP Requests (includes consumption tax).2,565 Requests0.01
Hong Kong and Singapore
  $0.190 per GB – first 10 TB / month data transfer out0.102 GB0.02
  $0.0090 per 10,000 HTTP Requests5,102 Requests0.01

Amazon Elastic Compute Cloud

US East (Northern Virginia) Region
  Amazon EC2 running Linux/UNIX
  $0.02 per Micro Instance (t1.micro) instance-hour (or partial hour)744 Hrs14.88
  Amazon EC2 EBS
  $0.10 per GB-month of provisioned storage18.293 GB-Mo1.83
  $0.10 per 1 million I/O requests30,188,451 IOs3.02
  $0.15 per GB-Month of snapshot data stored1.898 GB-Mo0.28
  $0.01 per 1,000 puts (when saving a snapshot)1,822 Requests0.02
  Amazon CloudWatch
  $0.00 per alarm-month – first 10 alarms0.141 Alarms0.00

Amazon Simple Notification Service

US East (Northern Virginia) Region
  First 100,000 Amazon SNS API Requests per month are free70 Requests0.00
  First 1,000 Amazon SNS Email/Email-JSON Notifications per month are free55 Notifications0.00

Amazon Simple Storage Service

US Standard Region
  $0.140 per GB – first 1 TB / month of storage used0.363 GB-Mo0.05
  $0.01 per 1,000 PUT, COPY, POST, or LIST requests54,515 Requests0.55
  $0.01 per 10,000 GET and all other requests219,232 Requests0.22
EU (Ireland) Region
  $0.01 per 1,000 PUT, COPY, POST, or LIST requests1 Requests0.01
  $0.01 per 10,000 GET and all other requests5 Requests0.01
Amazon Virtual Private Cloud


AWS Data Transfer (excluding Amazon CloudFront)

US East (Northern Virginia) and US Standard Regions
  $0.100 per GB – data transfer in per month5.169 GB0.52
  $0.000 per GB – first 1 GB of data transferred out per month1.000 GB0.00
  $0.150 per GB – up to 10 TB / month data transfer out8.804 GB1.32
  $0.010 per GB – regional data transfer – in/out/between EC2 Avail Zones or when using public/elastic IP addresses or ELB1.784 GB0.02
EU (Ireland) Region
  $0.000 per GB – first 1 GB of data transferred out per month0.000002 GB0.00

Total Charges due on February 1, 2011$23.60

January 2011 Bill for DLTJ on Amazon Web Services

Given the success and satisfaction level with this project on Amazon Web Services, I’m going to commit to a year-long “reserved” EC2 instance that drops the hourly cost for the instance even further. Based on running an instance continually for 732 hours per month, the cost for the EC2 part of the bill drops from $14.64 to $9.62. (The monthly total of $9.62 comes from $5.12 per month for the hourly charge plus one-twelfth of the annual $54 charge to reserve an instance.) Assuming the rest of the variables stay constant, that is $18.58 per month or about $220 per year.

Thursday Threads: Disruption in Library Acquisitions, Publishing, and Remedial Education plus Checking Assumptions of Cloud Computing and a National Digital Library

If it is Thursday it must mean it is time for another in this series of Thursday Threads posts. This week there are an abundance of things that could fall into the category of “disruptive innovation” in libraries and higher education. If you find these interesting, you might want to subscribe to my FriendFeed stream where these topics and more are posted and discussed throughout the week.
Continue reading

At the Intersection of the OCLC Records Use Policy and the WorldCat Local Cloud-based Library Management Service

Last Friday, Andrew Pace (Executive Director of Networked Library Services for OCLC) was interviewed by Richard Wallis of Talis on OCLC’s recent announcement of a cloud-based library management service. As part of that conversation, Richard and Andrew touched on the ongoing debate on the OCLC record use policy. Below is a transcript from that part of the interview (with time markers from the start of the interview).

Richard Wallis (27:00)
What about [libraries’] local data? By providing data up onto the OCLC platform, will that data be restricted in its use — how they can use it — or will it be totally open for them to use it in any way that they want to?
Andrew Pace (27:17)
That data is the library’s data.
Richard Wallis (27:21)
One of the reasons I ask that question is obviously we’re aware of the issues about bibliographic record reuse licensing that is going on at the moment. Do you see that conversation having any impact on the back-end data or the usage statistics data or anything like that?
Andrew Pace (27:41)
I imagine there will be service-level agreements we’ll build for the data that are going into library management services, but I am reluctant to combine what is going there with the record use policy discussions. I think as the record use policies are under revision — they are under discussion right now — I think libraries are cognizant of all of the data issues, but I am reluctant to tie the two together completely.
Richard Wallis (28:11)
So you probably see the bibliographic conversation separate from the raw data type conversation.
Andrew Pace (28:20)
Yeah, I think they are related to each other but I think they are separate conversations.

I think this is absolutely the right answer, and I’m glad to see the distinction between the shared bibliographic data and the holdings/circulation-transaction data so cleanly separated. They are related, but in the case of the former it is truly the library’s data. Earlier on in the interview, Andrew addressed the issue of how OCLC would respond to disclosure requests from law enforcement agencies.

Richard Wallis (21:35)
How would OCLC handle an inquiry under the [USA] Patriot Act or something like that?
Andrew Pace (21:42)
I might beg off on that as being a legal question, but it is one that we have asked about what it means for that data. I’m not sure it is going to be entirely different than how libraries would have to deal with it on a local system.
Richard Wallis (22:02)
I suppose the only concern is if you have the records for transactions in a signifcant number of libraries, it may actually be somewhere that government star people might want to wander and ask questions. I suppose that is where it is different in this environment.
Andrew Pace (22:20)
Yeah, and what I’m arguing is that it would be similar situation to software as a service or other hosted applications as well. But I’m not going to attempt any kind of legal answer since I’m not a lawyer.
Richard Wallis (22:37)
Ah, that’s disappointing. I could have quoted you back to yourself in a year’s time, but never mind, I understand why you ducked that question. I would have done so as well.

For me, this is further evidence that OCLC would consider the transaction data to be owned by the library. I’m not a lawyer either, but this would seem to push the responsibility for responding to a law enforcement agent request to the member library. Hopefully, there is legal precedent to make that stick.

Overall, it is a good interview that really puts some added definition to the plans for WorldCat Local Library Management Services.

The text was modified to update a link from to on August 27th, 2012.

OCLC’s WorldCat Local “Quick Start”

Last week, OCLC announced a “strategy to move library management services to Web scale.” With this move, OCLC is rebranding “WorldCat Local” to include functions typically associated with an integrated library system. From the press release:

OCLC plans to release Web-scale delivery and circulation, print and electronic acquisitions, and license management components to WorldCat Local, continuing the integration of library management services to create the Web-scale, cooperative library service. OCLC will begin piloting the Web-scale management service components this year.

There are many thing going on here. So, in order to get our bearings I think it is useful to break this down into two parts: the broader scope of web-scale integrated library system components and the narrower scope of the “quick start” trial version of the patron interface to But first we need to tackle the phrase “WorldCat Local.”

What is “WorldCat Local”?

Until now, many of us probably thought of WorldCat Local as a locally-scoped version of the database. It is, in fact, interesting to go back to the 11 April 2007 press release announcing the pilot of WorldCat Local; it says “the service will provide libraries the ability to search the entire WorldCat database and present results beginning with items most accessible to the patron.” Further on in the same release, it says “WorldCat Local service interoperates with locally maintained services like circulation, resource sharing and resolution to full text to create a seamless experience for the end user.”

OCLC now seems to be rebranding “WorldCat Local” from a discovery layer service to a replacement of a local integrated library system (although OCLC seems to go through great pains not to use the phrase “integrated library system” to describe this new offering). This rebranding varies depending on the page you are viewing: the press release has the quote at the top of this post while the first link in that press release (to the Web-scale, cooperative library management service) seems to speak of WorldCat Local in its 2007 definition. Andrew Pace’s blog post is somewhere in the middle: “OCLC is extending the WorldCat Local platform to include circulation and delivery, print and electronic acquisitions, and license management components.” In a conversation on the Code4Lib IRC channel, Roy Tennant says “AFAIK (as far as I know) WorldCat Local, which formerly was completely bounded by the catalog part, now will have different components, one of which is the catalog.”

Since the branding seems to be in flux, in this post I’m going to refer to WorldCat Local Web-scale Management Service (WCL-WMS) and WorldCat Local Discovery Service (WCL-DS). Any resemblance to current or future OCLC product names is quite definitely a coincidence.

Web-Scale Integrated Library System

Starting with WCL-WMS. In places on the OCLC website, this effort also goes by the name of Web-scale management services. OCLC describes its broad efforts this way:

OCLC’s vision is similar to Software as a Service (SaaS)1 but is distinguished by the cooperative “network effect” of all libraries using the same, shared hardware, services and data, rather than the alternative model of hosting hardware and software on behalf of individual libraries. Libraries would subscribe to Web-scale management services that include modular management functionality. Moreover, libraries would benefit from the network-level integration of numerous services that are not currently part traditional integrated library systems, e.g., Knowledge Base Integration, WorldCat Collection Analysis, WorldCat Selection, WorldCat Local, etc..

Further down the Web-scale management services page there is a list of proposed services: Web-Scale Circulation and Delivery, Web-Scale Print and Electronic Acquisitions, Web-Scale License Management, Web-Scale Self-Configuration, Web-Scale Workflow, and Web-Scale Cooperative Intelligence. The list also has brief descriptions of each of these services.

Trial Version of Patron Interface

The second component is a feature-limited version of the (formerly branded?) WorldCat Local interface. According to OCLC, the “quick start” version is missing the ability to have a consortial scope of holdings (between locally-owned and the-world), branch-level scoping within a local library, and local information display (such as notes and local URLs). The “quick start” version also won’t have the forthcoming metasearch integration or the ability to place holds via NCIP. (Holds can be place by redirecting the user to the home ILS and using whatever functionality is in place locally.) WCL-DS will pull availability information from local systems for items in which the library’s OCLC symbol is recorded in the WorldCat database. Status information is pulled either by Z39.50 look-up (in the case of Ex Libris Aleph systems) or HTML-scraping the Web OPAC display (all other systems). By offering it for free, OCLC is giving libraries the option of trying out a large subset of the WCL-DS functionality.

Initial Thoughts

From an OCLC-as-business perspective, this is a shrewd move. They may have effectively frozen two markets in one press release: libraries looking at discovery layers and libraries looking for new automation systems. For the former, the ability for libraries that already subscribe to OCLC FirstSearch — an interface that in transition to — to get a version of WCL-DS can be seen as a “gateway drug” that allows them to try out a thorough and robust challenger to Primo, Encore, Summon, and EBSCO Discovery Service, which is to say nothing about the open source contenders out there like Blacklight, VuFind, and the Extensible Catalog Project. In the case of the latter, there isn’t much movement in the ILS marketplace, but there is the 2-year-old Ex Libris URM and the soon-to-be-released report from the Open Library Environment project. (Will OCLC’s announcement of work in this area freeze out potential build partners for the OLE Project? The Web-scale document lists a feature of “A Service-Oriented Architecture (SOA) for interoperability with local environments and 3rd party business process systems (e.g., financial management, HR systems, and course management)” — one of the hallmarks of the OLE project.)

It is this second effect that has caught my eye the most. When pushing the back-room library automation tasks of circulation, acquisitions and such from local systems to “the cloud”, we’re fundamentally talking about a system migration. We’ve got to unload all of the records from our current system, massage them however we need to, put put them in another system. The twist comes from the fact that the other system is not the library’s own. I have no doubt that OCLC can work with existing ILS vendors to migrate their information into OCLC. My questions and concerns come at the other end — what if you want to migrate from the cloud back to a local system?

This is where considerations of OCLC-as-a-cooperative take over. If OCLC were acting as a cooperative — acting on the best behalf of its members — there would be no question that the data could easily come back out. The announcement, backlash, and response to the proposed changes to the record use policy are, however, a dark cloud2 over any reading of OCLC’s intentions. In a world where I could trust the cooperative to provide single-item lookups and batch aggregated extracts of holdings and acquisition data from the SaaS, what OCLC is offering could be a good thing. If it came along with demonstrations of real cost reductions (not simply “reducing the rate of rise” of costs) it would be even better. But OCLC has turned into a behemoth much larger than the collective whole of the collaborative, and I’m not sure I can trust them (collectively) at this point farther than I can throw the headquarters building in Dublin, OH.

Other Interesting Tidbits

The initial article in Library Journal by Marshall Breeding says: “OCLC said it will work with the more than 1,000 libraries and partners that are currently using OCLC library management systems in Europe and Asia Pacific to help build the new service.” In this context, I think “New Service” refers to WCL-WMS, and given the description is probably in reference to the OCLC PICA division.

In a future post, I’ll start to summarize the reactions to the OCLC announcement.


  1. Software as a Service is a model for deploying software where a firm provides the hosting and support of a system for a client. For example, Google Docs is a SaaS deployment for office automation applications and is a supplier of a customer relationship management system. As opposed to a model of software-purchase-plus-maintenance, SaaS involves a ongoing subscription fee. For more information, see the Wikipedia article on the topic. []
  2. Please pardon the cloud-computing pun. []