Purchasing Department August 16, 2012 RFP-MM-1213-11067 Enterprise Storage Addendum 1 A. SPECIFICATION CLARIFICATIONS / REVISIONS NONE B. REQUESTS FOR INFORMATION Oracle: 1) What version of Oracle is in the environment? a. Vers. 11 and 12 2) How many databases are in the Oracle environment? a. 35 14 are production 3) How many instances exist - is there a prod/dev/qa, etc. environment? a. 4 major applications with production, development test, and demo; 2 lesser applications with production, development, and test; and several with just production. There are also one-off copies of databases that get created for testing or recovery. 4) What are the sizes of the databases? a. 2.25TB in production databases, 3.28TB in Flash Recovery shared amongst all 35 databases, with an additional 2.37TB for other Oracle related storage 5) What is the workload - IO Read (Sequential & Random) and IO Write (Sequential & Random). a. A full blend of R/W loads and types with Random being the more predominant and a slight edge to more Read than Write. IOPS is not being tracked except from a combined storage throughput so database stats are not as accurate. 6) Is Oracle virtualized, if so what hypervisor? a. No 7) If Oracle is not virtualized, will it be or will it stay bare-metal? a. It will likely stay on physical servers, but we have no policy restricting placement.
NAS Capacity: 1) RFP states 30TB usable for NAS workloads - what applications will be provided via file storage? a. Because our environment has the potential to shift, some resources that are currently block delivered could change to file delivered and vice versa. Currently, we have the following: 1) File storage in Windows 7 and Server 2008 R2 environment including Home Drives and Common storage 2) nfs storage for Blackboard content system 3) nfs storage for VMware infrastructure 2) What are the application workloads - IO Read (Sequential & Random) and IO Write (Sequential & Random) for each application. 1) Windows storage is primarily Read with low IOPS 2) Blackboard content is primarily Read with bursts of Write and fairly low IOPS 3) VM storage is low IOPS Write with few, high IOPS Read 3) Is deduplication a requirement? a. No, but it will affect overall price and is desired by the college. 4) Is the 30 TB before deduplication algorithms? a. Because deduplication is unpredictable, yes, the 30 TB is before dedupe. 5) What deduplication ratio are you currently seeing in the NetApp Environment? 1) Windows storage is deduplicated at about 20% savings 2) Blackboard content is around 18% 3) VM storage is at least 50%, some volumes as high as 65% 6) Can deduplication be factored in when scoping the workload for 30 TB usable capacity? a. See questions 4 and 5 above SAN Capacity: 1) RFP states 20TB usable for SAN workloads - what applications will be provided via block storage? a. Because our environment has the potential to shift, some resources that are currently block delivered could change to file delivered and vice versa. Currently, we have the following: 1) iscsi storage for college video content 2) iscsi storage for video surveillance archives 3) iscsi storage for application delivery
2) What are the application workloads - IO Read (Sequential & Random) and IO Write (Sequential & Random) for each application. a. All current block level storage (iscsi) is fairly low IOPS 1) college video content is primarily Read 2) video surveillance archives is primarily Write 3) application delivery is primarily Read 3) Is deduplication a requirement a. No, but it will affect overall price and is desired by the college. 4) Is the 20 TB before deduplication algorithms? a. Because deduplication is unpredictable, yes, the 20 TB is before dedupe. 5) What deduplication ratio are you currently seeing in the NetApp Environment a. Some of the iscsi storage is not on the NetApp and none of it is deduplicated 6) Can deduplication be factored in when scoping the workload for 20 TB usable capacity? a. See questions 4 and 5 above Backup & Recovery/DR: 1) How many snapshots will be kept at one point in time? a. This varies based on application and environment (test vs. production). For example, home drives have a point-in-time reference generated every 2 hours (keeping 7 current), every day (keeping 3 current) and weekly (keeping 2 current). The database environment has a different schedule with additional snapshots converted to volumes for the test and development databases. 2) What backup software will be utilized for the environment? a. IBM Tivoli Storage Manager 3) What applications will be replicated to the DR location? a. The only thing planned so far is Oracle (Dataguard) and VMware (possible SRM) 4) What are the RPO/RTO requirements for each application? a. Has not been determined 5) What availability is required within the primary data center and also for DR? (99.999%?) a. Has not been determined Support: 1) How many years should be provided for manufacture software and support? a. 5 Years 2) Should an install quote be provided? Will this be a turnkey install? a. Installation is to be included. It will be beneficial to have installation cost/services broken out as a separate item.
Maintenance: 1) Does GRCC have a maintenance period for upgrades such as OS, Firmware, etc.? If so, how much time is granted and when? a. No Network: 1) Will the Broadcom FC Switches be replaced or maintained? If replaced is that a part of this RFP? a. Eliminated. The current FC environment is tied to the Fabric Attached portion of our Metro Cluster. The plan is to not reuse this setup. 2) Does GRCC currently have 10Gb network access? If so how many ports, if not is that a part of this RFP? a. Yes, 3) What switches are currently used for providing CIFS/NFS? a. All Servers: 1) What servers are currently utilized? a. IBM xseries and Cisco UCS 2) Please provide a list of HBAs and Firmware currently utilized in the servers for block based storage. a. None for the IBM (iscsi) and 6248 interconnect w/ 5.0(3)N2(2.02q) firmware for the UCS (vhbas). 3) Please provide a list of OS types and versions utilizing block based storage. a. Currently, Windows Server 2008 R2, Red Hat EL 4 VMware ESXi 5 4) Please provide a network diagram of the infrastructure. a. This is not part of this RFP process 5) Please provide the number of VMware hosts, their consolidation ratios, versions, and HBAs/NICs utilized for Block & File storage access a. We have 18 ESX hosts running 3, 4, and 5 versions 6) Please provide PerfMon/PerfStat data for each of the applications running in the environment that will be placed on the new storage array(s). a. This data is not available at this time Misc.: 1) Do you want File and Block data to be serviced from a unified appliance? a. The file and block can be serviced by a unified appliance or two different solutions 2) Section 5.1, first bullet, calls for 30TB available NAS capacity. How much of the 30TB will be used for the Vmware environment? a. We are currently using 5TB for VMware storage in our environment, but we are open to moving it to the block storage
3) Would you prefer to use NFS in your Vmware environment for ease of migration? a. The migration will be done with tools, either to a new NFS storage or block level. Both will take the same effort. ( No Preference ) 4) What are your future growth and scalability requirements for the proposed new storage? Currently you list 30TB of usable NAS capacity and 20TB of usable SAN capacity--how much additional NAS and SAN capacity should the proposed storage be able to scale up to in the future? a) The requested storage is based off of a 3yr projection and we are unable to project farther into the future. At the end of the 3 years, technology would help dictate future expansion and growth. The NAS capacity should at least provide up to 60TB in the future. 5) Please provide information on server hosts attaching to both the NAS and SAN storage, including number of hosts, operating systems, and application information. a) We have 13 VMware ESX hosts attached, 10 Linux systems running Oracle, various Windows 2003 2008 R2 and Linux systems attached running utility type applications or user data. 6) If we propose storage models that have multiple RAID level configuration capability (RAID 10, 5, 6, etc.), what are the RAID requirements for the 30TB of usable NAS capacity and 20TB of usable SAN capacity defined in the RFP? Please specify different RAID levels for different tiers of capacities as appropriate (e.g.. 10GB of RAID 10 usable on NAS / 20GB of RAID 5 usable on NAS). Alternatively, please provide performance characteristics required for the usable NAS and SAN storage including IOPS and throughput information as well as percentage read/write characteristics. Please include information on growth and scalability if known for the different tiers. Related to above, are there lower tiers of capacity with lesser performance requirements that would be suited to lower-cost SATA disk capacity versus higher cost SAS and FC? a) Please provide the best configuration for the solution and explain why it is the best. The recipient of the RFP award will have the opportunity to go over the environment in detail before the purchase is made. This will ensure the best solution for the college but allow flexibility in the responses.
7) Please define performance requirements (IOPS/throughput, percentage read/write characteristics) for all NAS and SAN storage. If there are collected perfmon or iostat data files for the existing storage environment relevant to sizing the new storage environment, please include. Please include any information on existing issues with bottlenecks or I/O-related performance. What is the user feedback on application and database performance in the current environment? a) We currently do not have any other information than what was posted on the previous questions and answers. 8) In the RFP, it states you currently have quantity 4 Broadcom E200 FC switches. Did you actually mean Brocade E200 switches (16-port 4Gbps switches)? Broadcom does not have an E200 model, but Brocade does. a) Brocade is correct. 9) We want to determine the number of ports needed to replace the existing SAN fiber channel switches. How many ports are you currently using for server hosts, how many for the SAN storage devices, and how many for ISL links? Is there a requirement for additional ports to be used for future server hosts or other SAN-connected devices and if so how many? a) We currently do not use FC to the host, only iscsi. The existing FC ports are used in the current deployed solution and would not be re-used. If a solution is selected that uses FC connections, we would determine the correct port count prior to the purchase 10) RFP requests information on how the proposed solution can replicate to a remote site over a 1Gb/s connection. What is the distance between the primary site to the planned remote site? a) A proposed location is 5 miles from our primary 11) Are implementation services and training to be included in the RFP response in addition to the hardware? How do you plan to handle data and database migration from the existing storage to the new storage--will you require services for this as well? Proposals for migration services would require additional information, such as database information and requirements for planned outage versus a live migration. a) Implementation services are not required to be included in the RFP response, but an option for block of time or estimates can be submitted and options. 12) What does the infrastructure look like between the two sites bandwidth, latency, distance. a) We are currently attached via fiber running 10Gb Ethernet
13) RFP says its FC mesh, is that FC end to end, or is FCIP in the mix? a) FC is only in the existing solution, no FCIP 14) If FCIP is in the mix, is the Ethernet infrastructure bandwidth dedicated to the existing metro cluster, or is it shared. If shared, is QOS in place? If so, what does that look like. a) No FCIP 15) What is the life expectancy of the replacement solution? a) forever would be nice, but at least 5 years 16) How large does this environment need to scale in the next few years? a) The RFP is based on a 3 year forecast, but we would need the NAS to expand to 60 TB 17) Is there any netapp autosupport data for us to look at to understand what the workload on these assets is? a) No 18) Does GRCC Need a new rack(s) for this solution? a) That would depend on the solution, we currently have multiple ½ rack spaces available. PDU, Cable Management, UPS b) No UPS or cable management, PDU would depend on the power requirements of the solution. You can provide them as optional cost or options 19) 10GB or 1GB connections for NAS Front End connectivity to the LAN, Fiber or Copper? a) Fiber 10GB would be preferred 20) Are there any bare metal SQL or other application instances that need to integrate with Microsoft VSS for Block Storage? a) No 21) Does the DR Site need to be an exact replica? a) The DR site is a future deployment, the response should just provide the capability of the solution to implement in the future Does all data need to be replicated? b) No i.e. All 50TB or just 45 out of 80 volumes? c) We do not have a break down, future project
Both NAS and block? Do NAS shares need to be accessed in a DR scenario? d) Not determined yet, but I believe yes. 22) Professional Services, do the following services need to be provided Please answer yes or no? If yes please provide a brief description of the services to be included. If allowances have been established, please make note of those amounts: a) Installation Services: As an option i. Physical install into rack ii. Cable Management/Labeling iii. Power Cabling b) Migration Services No i. # of Windows server volumes to be migrated? ii. # of VMware servers to be migrated? c) Also please make note of any Server with RDMs i. # of Linux server volumes to be migrated? ii. # of Unix server volumes to be migrated? iii. # of CIFS/NFS Shares? d) Training Services As an option i. # of daily/basic administrators to be trained 1 ii. # of advanced administrators to be trained 2 23) Is there a word document available of the RFP? a) Yes, please see attached. 24) How many CIFS users should we size the NAS for? a) Concurrent users ( 2,000 ): Servers ( 5 ) 25) Are there any performance sizing requirements for the Block and File Storage? IOPs, Throughput? a) Only what was provided in previous answers Please acknowledge receipt of this addendum with you bid submission.